Every shredder I've seen for the last decade has been a crosscut shredder instead of the old style.
The shredded paper strips look like what comes out of the low-end Champion shredder I bought at Office Depot last year. Including the slightly serrated edges. That thing just cuts paper into 8mm strips. As a security device, it's not much.
(I bought one to use as a paper slitter to make 8mm paper tape to be printed on by antique Model 14 Teletypes. It's not a great paper slitter, but running adding machine rolls though it made enough tape to get the Teletypes working.)
People started thinking that they could build silicon to do things even faster, and thus the ASIC market started to emerge and take off.
Well, maybe. Like much in the Bitcoin world, some of this is a scam. At least one of the "ASIC" products turned out to be an FPGA. As of right now, it's not clear that anyone is actually shipping an ASIC-based Bitcoin mining device. Suckers can pre-order from either of twovendors. Payment is in US dollars, not Bitcoins.
That's what most "apps" really are. Most of them don't do much more than a web page could. But they put the content owner firmly in control of the user experience.
"Do not adjust your television. We are controlling it. We control the horizontal. We control the vertical..."
I'm surprised the NTSB wanted something as fragile as an iPhone. I would have expected them to go for something that had a ruggedized, waterproof model in the product family.
Rugged smartphones have been around for a while, but in 2012, they got bigger screens and current electronics. The Samsung Galaxy Rugby Pro, the Honeywell Dolphin 70e, the rather bulky Caterpillar B10 Smartphone, and the thin Nautiz X1 all meet basic military ruggedization standards while running reasonably current Android versions.
That means they know exactly where it was launched from too.
No, not for a rocket. That works for artillery, where the trajectory is purely ballistic. The U.S. Army has had good counter-battery fire systems for decades. The Fire Finder performance standard is that fire should be returned accurately within one minute.
For a rocket, you need an airborne radar that can see the launch, and something that can shoot back at the launch site.
Hamas is using unguided rockets, similar to WWII
Katyushas. Those are capable of hitting a city, but are not accurate enough to hit an airfield, a hangar, or a military unit. So they make the Israelis angry but don't win battles.
Some new models fail in the marketplace. GM's Bob Lutz, in "Car Guys vs. Bean Counters", has a lot to say about how that happens in the auto industry. Sometimes a product bombs, and entire huge plants with hundreds of millions in tooling have to be scrapped.
Microsoft isn't used to having that happen. It's been so dominant that it could impose a new model on its user base. But with Microsoft's tiny market share in mobile, it can't do that there. Trying to push a failed model in mobile onto the desktop is backfiring.
Well, what do you expect with a system-on-a-chip? A modern high-volume consumer product should have one IC. That's the whole point of SOIC. It's a bit hard for phones, because they have all those radios that need some isolation, but a modern game console ought to have a very low parts count. Makes assembly very cheap, too.
The legal arguments are interesting. It's amusing to see lawyers struggle with reasoning through analogy. They're trying to hammer property law, trespass law and assault law into covering this, and it's not working.
In almost all modern online attacks, the immediate source of the the attack is a machine owned by an innocent third party. While this is common online, it is a rare situation in the physical world. It can come up in auto repossessions where the repossession was not legally authorized, the repossession agent reasonably believed that it was, and the vehicle owner resisted. Most states have specific laws in that area, and repossession agents are limited in what they can do.
Trust me, everybody would loooooooove for the computer to take instructions like a human but it's not going to happen because of everything that's implicitly understood. So you can teach this computer to fold a shirt, if you hand it an XS shirt and an XXL shirt will it figure out that it must adapt the folding action to the size of the shirt?
Yes, that's the hard problem in learning from demonstration - working back from the demonstration to a model which can be generalized to new tasks. One way to approach this is by doing the same task with variations - guide the robot through folding various different shirts, and then use a machine learning system to separate the commonalities from the differences. There's been some progress in recent years in making this work. It's not very powerful yet, but it's getting to be good enough for teaching assembly line robots.
Yes, there are the CS people doing actual science, performing studies, creating the new stuff for all the future SE's to use, but they're the vast minority. If you want to see more than 2 in the same room go to SIGGRAPH.
True. More true in the 1990s, when rendering and physical simulation were being figured out. Game development used to need theoreticians. Now it needs people who can wrangle the large number of people and vast amounts of data that go into an A title.
Wikipedia editing is not about formatting. It's not about font or size. The markup language includes links and many macros with specific parameters. Those are where users require assistance. {{cite book}}, for example, requires many parameters which could be filled in automatically, such as author, title, ISBN, publisher, and date of publication.
Wikipedia's big structure problem, though, is that it is a wiki only. Some kinds of information belong in a different kind of structure. Items like "State Route 92" belong in a spatial database tied to maps. Music and bands belong in databases where songs and performers are automatically indexed. Wikipedia is full of manually maintained popular culture related list items which should be generated with a SELECT statement.
On the actual site (quit linking to some blog that links to a site of interest, Slashdot submitters), it's just typical jerks, mostly in high school. Somebody at the Secret Service will have to read through all that dreck looking for someone who might conceivably be a threat. It's mostly just kids mouthing off. Ones who are both making threats and have guns may get some attention.
It's sad reading the Facebook pages of some of those people. Their future is dim. For white people with only a high school education, life in the US has become slightly worse each year since 1973. For them, there is no American Dream. Hard work is no longer enough. Of course they're angry. The GOP and Fox News exploit and direct that anger at Obama, but they didn't create it.
This seems to have the same problem as shape-memory alloys. Those change shape quickly when heated above their transition temperature, but the amount of energy you have to put in is far more than you get out. Then they have to cool down before they can be cycled again. Power to weight ratio is good, but energy to weight ratio is poor because the cycle time is slow.
Probably not all that useful as a general actuator.
I see a future for these boxes though. Up the price and sell it as a security and monitoring device.
That's been tried at the ISP level. There are a few Christian ISPs that filter out stuff they don't like. There are Jewish ISPs which filter out stuff they don't like. There's an ultra-Orthodox Jewish ISP which filters everything except a very short list of approved sites. (Here's such a list.)
They're all very small operations. Self-censorship is a tiny business. Most of the people clamoring for web censorship want to censor the Internet for other people.
Watched the video. It's all about their little hardware box (which is some ARM machine), and says nothing about how it blocks ads. At the wire level, you can certainly apply a domain blocklist, for which there are already many free software tools. That gets rid of many ads, but not all of them.
Some (not yet many) sites resist ad blocking.
Some Flash-driven videos won't play if you block their ad server. Some get the ads and the video from the same place. Some ad services have each site create a subdomain (like "ads.example.com") for ad serving, so blocking by second level domain doesn't work. Look at the constantly changing blocklists for AdBlock. The problem is almost as bad as signature-based virus detection. The people with this little box say nothing about this.
The one big advantage this device offers is the ability to block ads on closed systems like Apple products. A big disadvantage is that the device has a backdoor into your data stream and could be an attack vector for eavesdropping.
Many heavy industries have cooling ponds. The problems are routine and known.
I once worked for a company near Detroit which made heavy hydraulic equipment. The R&D operation had a building full of test cells in which locomotive transmissions and similar big stuff was exercised for months on end. The dummy loads for the machines dumped the energy into water. The water went to a cooling pond in front of the plant. The pond was made to look like a large decorative fountain with water sprays. But it was really a heat sink.
There's a shortage of airline pilots because the job doesn't pay well any more and takes extensive training. Training most US airlines are not willing to pay for. The WSJ is whining that the FAA raised the standards for an Airline Transport Pilot rating and requires pilots to get more sleep. That's in response to the crash of Continental flight 3407 on February 12, 2009. The WSJ conveniently does not mention that.
Some airlines do pay for training. Here's the British Airways training program. BA pays pilot trainees as employees through the whole training process. Most US airlines expect pilots to work for years for less than a typical city bus driver makes to build up their hours before they fly the big iron.
There many good "organic" modelers. Autodesk Mudbox is widely used by pros. Curved surface volumetric modellers go back a long way. I used one of the very first back in the 1980s, one based on deformable superellipsoids and running on a Symbolics LISP machine.
As for the "physical Turing test", if your demo reel doesn't show that you can pass that, it won't get you in the door at Pixar.
This guy is not a security expert. His bio:
"Before joining the Council, I was a sell-side commodity strategist at Louis Capital Markets." That's a brokerage firm. A "sell-side analyst" is really a PR guy who generates happy-talk "buy" recommendations which are sent to customers.
I go to occasional VC conferences in Silicon Valley, and get to hear what the VCs are doing. It's not looking good. The VC industry as a whole hasn't made money since 2001 or so. During the dot-com boom, many new venture funds were created, resulting in an influx of dumb money. Most of them haven't been profitable.
The problem is this. Before the dot-com boom, venture capitalists usually funded a startup which was going to make something. So they'd fund a few engineers for a few years, and about 1 time in 10, something good would come out that paid for the unsuccessful tries. This was a good business model and it drove Silicon Valley.
Dot-com startups weren't about technology. They were about marketing and market share. So they had to be funded beyond the R&D phase, well into the growth phase, before the winners and losers became clear. The loss per failure was much higher than when VCs were involved in technology startups.
This model has persisted in the post dot-com era and into the "Web 2.0" era. Most of the ideas one sees at VC meetings are minor variations on popular ideas. I've seen a presentation for a social network for cats. Some innovative technologies are proposed, but often they're not big wins.
About 1 in 10 VC-funded companies makes it big. 2 to 3 in 10 go bust. The rest end up in "zombie mode" - they generate enough cash to pay their expenses, but can't pay back their investors. A big headache in the VC industry is dealing with the growing army of zombies. They're more profitable alive than dead, so they're not killed off, but they're a net loss. There are a lot of half-dead "social" startups around. Tech startups tended to be sold off for the technology or shut down. That's what VCs mean by "lack of an exit".
Protecting Manhattan isn't that difficult. It's clear that the Con Ed station on 14th St needs to be raised; that's too important to be flooded out again. The subway system needs flood gates at several points. The London and Singapore systems have flood gates. The old Pennsylvania Railroad North Tunnels have flood gates, which Amtrak didn't maintain and were supposed to be fixed after 2001 as an anti-terrorism measure.
Some of the subway stations need extra protection, especially South Ferry. They need strong emergency flood barriers. Sandbags didn't work because a big piece of wood (about 1' x 1' by 15') from a construction site crashed through them and ended up in the booking hall. They need steel barriers that are raised out of the ground when necessary. Extra pumping capacity with backup power is indicated, too.
Those are no-brainers. After major hurricanes two years in a row, there's no question that those basic fixes are needed. Beyond that, it might be worthwhile to raise the ground level of the parks in the Battery Park area by a few meters. FDR Drive may need a flood wall south of the Brooklyn Bridge. Those are less urgent.
Barrier islands like Fire Island and the Rockaways, and the Jersey shore, are too low to fix. Just make sure everybody evacuates in time. (About 140 people refused to evacuate Fire Island, and getting them off after the island had been cut in two by the storm risked the lives of emergency personnel. The first group of rescuers had to be rescued.) Require Florida-level hurricane protection in house construction. Require paid-up private insurance for anyone who wants to build in the flood zone. Put in hurricane-resistant solar panel powered street lights (a commercially available product), so there's some light no matter what happens. A strict "no tall trees near power lines" policy may be necessary in the coastal zone.
New York State has a valuable resource - big rocks. Where roads and railroad tracks need to be protected against washouts, big rocks, too big for a storm to move (granite boulders the size of a SUV) should be used extensively.
(Forget the "balloon tunnel plug" idea. Something like that was used at the Penn Station yards, and it burst when hit by something.)
Slashdot Mirror
Every shredder I've seen for the last decade has been a crosscut shredder instead of the old style.
The shredded paper strips look like what comes out of the low-end Champion shredder I bought at Office Depot last year. Including the slightly serrated edges. That thing just cuts paper into 8mm strips. As a security device, it's not much.
(I bought one to use as a paper slitter to make 8mm paper tape to be printed on by antique Model 14 Teletypes. It's not a great paper slitter, but running adding machine rolls though it made enough tape to get the Teletypes working.)
People started thinking that they could build silicon to do things even faster, and thus the ASIC market started to emerge and take off.
Well, maybe. Like much in the Bitcoin world, some of this is a scam. At least one of the "ASIC" products turned out to be an FPGA. As of right now, it's not clear that anyone is actually shipping an ASIC-based Bitcoin mining device. Suckers can pre-order from either of two vendors. Payment is in US dollars, not Bitcoins.
That's what most "apps" really are. Most of them don't do much more than a web page could. But they put the content owner firmly in control of the user experience.
"Do not adjust your television. We are controlling it. We control the horizontal. We control the vertical..."
I'm surprised the NTSB wanted something as fragile as an iPhone. I would have expected them to go for something that had a ruggedized, waterproof model in the product family.
Rugged smartphones have been around for a while, but in 2012, they got bigger screens and current electronics. The Samsung Galaxy Rugby Pro, the Honeywell Dolphin 70e, the rather bulky Caterpillar B10 Smartphone, and the thin Nautiz X1 all meet basic military ruggedization standards while running reasonably current Android versions.
That means they know exactly where it was launched from too.
No, not for a rocket. That works for artillery, where the trajectory is purely ballistic. The U.S. Army has had good counter-battery fire systems for decades. The Fire Finder performance standard is that fire should be returned accurately within one minute. For a rocket, you need an airborne radar that can see the launch, and something that can shoot back at the launch site.
Hamas is using unguided rockets, similar to WWII Katyushas. Those are capable of hitting a city, but are not accurate enough to hit an airfield, a hangar, or a military unit. So they make the Israelis angry but don't win battles.
Some new models fail in the marketplace. GM's Bob Lutz, in "Car Guys vs. Bean Counters", has a lot to say about how that happens in the auto industry. Sometimes a product bombs, and entire huge plants with hundreds of millions in tooling have to be scrapped.
Microsoft isn't used to having that happen. It's been so dominant that it could impose a new model on its user base. But with Microsoft's tiny market share in mobile, it can't do that there. Trying to push a failed model in mobile onto the desktop is backfiring.
Well, what do you expect with a system-on-a-chip? A modern high-volume consumer product should have one IC. That's the whole point of SOIC. It's a bit hard for phones, because they have all those radios that need some isolation, but a modern game console ought to have a very low parts count. Makes assembly very cheap, too.
The legal arguments are interesting. It's amusing to see lawyers struggle with reasoning through analogy. They're trying to hammer property law, trespass law and assault law into covering this, and it's not working.
In almost all modern online attacks, the immediate source of the the attack is a machine owned by an innocent third party. While this is common online, it is a rare situation in the physical world. It can come up in auto repossessions where the repossession was not legally authorized, the repossession agent reasonably believed that it was, and the vehicle owner resisted. Most states have specific laws in that area, and repossession agents are limited in what they can do.
I still have some CSS like this:
Mozilla finally got with the program and implemented border-radius without a prefix. It took a while.
Trust me, everybody would loooooooove for the computer to take instructions like a human but it's not going to happen because of everything that's implicitly understood. So you can teach this computer to fold a shirt, if you hand it an XS shirt and an XXL shirt will it figure out that it must adapt the folding action to the size of the shirt?
Yes, that's the hard problem in learning from demonstration - working back from the demonstration to a model which can be generalized to new tasks. One way to approach this is by doing the same task with variations - guide the robot through folding various different shirts, and then use a machine learning system to separate the commonalities from the differences. There's been some progress in recent years in making this work. It's not very powerful yet, but it's getting to be good enough for teaching assembly line robots.
Yes, there are the CS people doing actual science, performing studies, creating the new stuff for all the future SE's to use, but they're the vast minority. If you want to see more than 2 in the same room go to SIGGRAPH.
True. More true in the 1990s, when rendering and physical simulation were being figured out. Game development used to need theoreticians. Now it needs people who can wrangle the large number of people and vast amounts of data that go into an A title.
Wikipedia editing is not about formatting. It's not about font or size. The markup language includes links and many macros with specific parameters. Those are where users require assistance. {{cite book}}, for example, requires many parameters which could be filled in automatically, such as author, title, ISBN, publisher, and date of publication.
Wikipedia's big structure problem, though, is that it is a wiki only. Some kinds of information belong in a different kind of structure. Items like "State Route 92" belong in a spatial database tied to maps. Music and bands belong in databases where songs and performers are automatically indexed. Wikipedia is full of manually maintained popular culture related list items which should be generated with a SELECT statement.
On the actual site (quit linking to some blog that links to a site of interest, Slashdot submitters), it's just typical jerks, mostly in high school. Somebody at the Secret Service will have to read through all that dreck looking for someone who might conceivably be a threat. It's mostly just kids mouthing off. Ones who are both making threats and have guns may get some attention.
It's sad reading the Facebook pages of some of those people. Their future is dim. For white people with only a high school education, life in the US has become slightly worse each year since 1973. For them, there is no American Dream. Hard work is no longer enough. Of course they're angry. The GOP and Fox News exploit and direct that anger at Obama, but they didn't create it.
This seems to have the same problem as shape-memory alloys. Those change shape quickly when heated above their transition temperature, but the amount of energy you have to put in is far more than you get out. Then they have to cool down before they can be cycled again. Power to weight ratio is good, but energy to weight ratio is poor because the cycle time is slow.
Probably not all that useful as a general actuator.
I see a future for these boxes though. Up the price and sell it as a security and monitoring device.
That's been tried at the ISP level. There are a few Christian ISPs that filter out stuff they don't like. There are Jewish ISPs which filter out stuff they don't like. There's an ultra-Orthodox Jewish ISP which filters everything except a very short list of approved sites. (Here's such a list.)
They're all very small operations. Self-censorship is a tiny business. Most of the people clamoring for web censorship want to censor the Internet for other people.
Watched the video. It's all about their little hardware box (which is some ARM machine), and says nothing about how it blocks ads. At the wire level, you can certainly apply a domain blocklist, for which there are already many free software tools. That gets rid of many ads, but not all of them.
Some (not yet many) sites resist ad blocking. Some Flash-driven videos won't play if you block their ad server. Some get the ads and the video from the same place. Some ad services have each site create a subdomain (like "ads.example.com") for ad serving, so blocking by second level domain doesn't work. Look at the constantly changing blocklists for AdBlock. The problem is almost as bad as signature-based virus detection. The people with this little box say nothing about this.
The one big advantage this device offers is the ability to block ads on closed systems like Apple products. A big disadvantage is that the device has a backdoor into your data stream and could be an attack vector for eavesdropping.
with their habit of creating walled gardens deliberately incompatible with competitors' platforms finally catching up to them.
Everybody from Apple to Comcast has a "walled garden" now. Even Canonical has an "app store". The New York times is thriving behind its paywall.
Many heavy industries have cooling ponds. The problems are routine and known.
I once worked for a company near Detroit which made heavy hydraulic equipment. The R&D operation had a building full of test cells in which locomotive transmissions and similar big stuff was exercised for months on end. The dummy loads for the machines dumped the energy into water. The water went to a cooling pond in front of the plant. The pond was made to look like a large decorative fountain with water sprays. But it was really a heat sink.
(It's gone; there's a mall there now.)
There's a shortage of airline pilots because the job doesn't pay well any more and takes extensive training. Training most US airlines are not willing to pay for. The WSJ is whining that the FAA raised the standards for an Airline Transport Pilot rating and requires pilots to get more sleep. That's in response to the crash of Continental flight 3407 on February 12, 2009. The WSJ conveniently does not mention that.
Some airlines do pay for training. Here's the British Airways training program. BA pays pilot trainees as employees through the whole training process. Most US airlines expect pilots to work for years for less than a typical city bus driver makes to build up their hours before they fly the big iron.
A First Officer (copilot) on RyanAir starts at $3700 a month.
Another ad for a Kickstarter campaign. Yawn.
There many good "organic" modelers. Autodesk Mudbox is widely used by pros. Curved surface volumetric modellers go back a long way. I used one of the very first back in the 1980s, one based on deformable superellipsoids and running on a Symbolics LISP machine.
As for the "physical Turing test", if your demo reel doesn't show that you can pass that, it won't get you in the door at Pixar.
Ivan Sutherland, still alive and working at 74. Wow.
This guy is not a security expert. His bio: "Before joining the Council, I was a sell-side commodity strategist at Louis Capital Markets." That's a brokerage firm. A "sell-side analyst" is really a PR guy who generates happy-talk "buy" recommendations which are sent to customers.
This has little to do with the recession.
I go to occasional VC conferences in Silicon Valley, and get to hear what the VCs are doing. It's not looking good. The VC industry as a whole hasn't made money since 2001 or so. During the dot-com boom, many new venture funds were created, resulting in an influx of dumb money. Most of them haven't been profitable.
The problem is this. Before the dot-com boom, venture capitalists usually funded a startup which was going to make something. So they'd fund a few engineers for a few years, and about 1 time in 10, something good would come out that paid for the unsuccessful tries. This was a good business model and it drove Silicon Valley.
Dot-com startups weren't about technology. They were about marketing and market share. So they had to be funded beyond the R&D phase, well into the growth phase, before the winners and losers became clear. The loss per failure was much higher than when VCs were involved in technology startups.
This model has persisted in the post dot-com era and into the "Web 2.0" era. Most of the ideas one sees at VC meetings are minor variations on popular ideas. I've seen a presentation for a social network for cats. Some innovative technologies are proposed, but often they're not big wins.
About 1 in 10 VC-funded companies makes it big. 2 to 3 in 10 go bust. The rest end up in "zombie mode" - they generate enough cash to pay their expenses, but can't pay back their investors. A big headache in the VC industry is dealing with the growing army of zombies. They're more profitable alive than dead, so they're not killed off, but they're a net loss. There are a lot of half-dead "social" startups around. Tech startups tended to be sold off for the technology or shut down. That's what VCs mean by "lack of an exit".
Will this be a "certified dumb enough for school use during tests" device?
Protecting Manhattan isn't that difficult. It's clear that the Con Ed station on 14th St needs to be raised; that's too important to be flooded out again. The subway system needs flood gates at several points. The London and Singapore systems have flood gates. The old Pennsylvania Railroad North Tunnels have flood gates, which Amtrak didn't maintain and were supposed to be fixed after 2001 as an anti-terrorism measure.
Some of the subway stations need extra protection, especially South Ferry. They need strong emergency flood barriers. Sandbags didn't work because a big piece of wood (about 1' x 1' by 15') from a construction site crashed through them and ended up in the booking hall. They need steel barriers that are raised out of the ground when necessary. Extra pumping capacity with backup power is indicated, too.
Those are no-brainers. After major hurricanes two years in a row, there's no question that those basic fixes are needed. Beyond that, it might be worthwhile to raise the ground level of the parks in the Battery Park area by a few meters. FDR Drive may need a flood wall south of the Brooklyn Bridge. Those are less urgent.
Barrier islands like Fire Island and the Rockaways, and the Jersey shore, are too low to fix. Just make sure everybody evacuates in time. (About 140 people refused to evacuate Fire Island, and getting them off after the island had been cut in two by the storm risked the lives of emergency personnel. The first group of rescuers had to be rescued.) Require Florida-level hurricane protection in house construction. Require paid-up private insurance for anyone who wants to build in the flood zone. Put in hurricane-resistant solar panel powered street lights (a commercially available product), so there's some light no matter what happens. A strict "no tall trees near power lines" policy may be necessary in the coastal zone.
New York State has a valuable resource - big rocks. Where roads and railroad tracks need to be protected against washouts, big rocks, too big for a storm to move (granite boulders the size of a SUV) should be used extensively.
(Forget the "balloon tunnel plug" idea. Something like that was used at the Penn Station yards, and it burst when hit by something.)