Right. Especially since the Computer Museum in Sunnyvale already has a videogame section. They have the original PDP-1 Spacewar, the original Pong game, most of the early game consoles, etc. They even have some of the early games playable in emulators.
I thought Google's algorithms were supposed to decide which results were the best and put those at the top. I thought the "+1" was supposed to be a link to a site from another authoritative site. Is this an admission that this mechanism doesn't work? That it's been hopelessly gamed by spammers?
Basically, yes.
It's surprising that Google did this, after their disastrous experience with counting Yelp and CitySearch recommendations to compute ratings in Google Places. Last October, Google merged Places results from the previously independent Google Maps system into main web search results. Google Places ranking is based on location and recommendations, both of which can be easily faked and spammed. Within a month, we were seeing SEO firms advertising "Guaranteed first page listings or your money back". Within two months, Google Places spam was dominating results in local service categories like "dry cleaners" and "locksmiths".
Then Google was humiliated by the New York Times investigation into the J.C. Penny spam job. That was classic link farming, something Google supposedly knew how to stop. Google penalized J.C. Penny, but that created a political backlash. Now Google had admitted that they manually adjusted search results. The European Union's antitrust authorities started investigating that.
Blekko started sniping at Google for being so inept at dealing with web spam. Blekko improved results just by manually blocking the usual suspects, like "about.com".
Demand Media went public and raised awareness of all the junk coming from "content farms". Google tried penalizing content farms, but it turned out that Google can't tell a content farm from a catalog shopping site, and was applying penalties to catalogs of real, buyable stuff. This angered some legitimate companies.
Google added a "mark as spam" link on search results a few weeks ago. (So they do have a "hate" button.) Now they've added a "like" button. They seem to be chasing after Blekko's "crowdsourced" search.
But it won't work. Blekko works only because nobody bothers to spam Blekko. They're too small to bother. Google's "like" mechanism will be spammed. Google can't stop people from creating massive numbers of phony Google accounts (see Jiffy Gmail Account Creator). As the previous Slashdot article points out, there are now far more social networking accounts than humans on the planet. Most of them were created by programs, not humans.
Really, though, Google has a "like" button because Facebook has a "like" button.
From the New Yorker article, here's the one big item of new info he's discovered:
In the standard historical accounts, the way that the bomb's gun mechanism worked was by shooting a cylindrical âoemaleâ uranium projectile into a concave, stationary uranium target. This act of atomic coitus created a mass sufficient to produce a critical reaction. The mass of the projectile was said to be 38.5 kilograms, and the mass of the target was said to be 25.6 kilograms. But no matter how many times Coster-Mullen did the math the numbers never quite worked out in a way that allowed the projectile and the target to fit inside the gun barrel while remaining subcritical.
The source of the error, Coster-Mullen recognized, was an assumption that every (male) researcher who studied the subject had made about the relation between projectile and target. These scholars had apparently been unable to conceive of an arrangement other than a "missionary position" bomb, in which a solid male projectile penetrated a vessel-like female target. But Coster-Mullen realized that a female-superior arrangement - in which a hollow projectile slammed down on top of a stationary cylinder of highly enriched uranium - yielded the correct size and mass.
Now that's a surprise. I wonder why the Manhattan Project did it that way, shooting the larger mass into the smaller mass. Maybe that was to get the assembly to hold together longer while the chain reaction initiated.
(For those of you who slept through the atomic weapons part of high school physics, the Hiroshima bomb was a "gun" bomb, where the tube from an artillery piece was used to fire one subcritical mass of uranium into another subcritical mass, producing a critical mass. That's been an obsolete technology for half a century, because such bombs are so bulky and require much more fissionable material than implosion bombs, but it works. The Trinity and Nagasaki bombs were implosion devices, of course.)
These frequent "big materials breakthrough" articles really should wait until they've been reviewed in some publication that knows something about the subject, like Chemical Engineering News. The paper, "Agro-waste nanocomposites for automotive applications", presented at the American Chemical Society is available. The claims there aren't as strong as the ones in the press release. Last year, the same author presented "Agro-Wastes Nanocomposites for Medical Application". Wonder what happened to that.
The trouble with many of these "new materials" is that they have some awful flaw. This one, for example, is "biodegradable". That means it rots. That's OK for packaging, but not for parts. Then there are basic questions, like will it tolerate water? Can it be made into thread, sheet, or film? Made at a reasonable cost?
There's been interest in finding useful things to make out of cellulose for the last century. There's so much agricultural waste around, and it would be nice to use it for something. Most of the ideas don't work out, but people keep trying.
France, Germany, and the US are all sending over rugged robots designed to work in disaster areas, collapsed building sites, and war zones. Strangely, Japan doesn't seem to have those.
I'm amazed that TEPCO hadn't at least brought in a few hobby-type R/C quadrotors or helicopters with TV cameras to get a look at areas they couldn't reach, like the spent fuel pools. For days, they didn't even know the water levels in those pools. Attempts were made to peer through holes in the roof with high-altitude flyovers, but that produced results like "we see some water in the pools, but can't tell how much."
The biggest problem with TEPCO has been eliminated. The CEO was just canned for "health reasons". He's been in hiding since the earthquake.
Just like you can't buy a car, reverse engineer it, and start mass-producing that car and competing with Mazda with their own product. You bought the car, not the plans to the car, nor the rights to sell it.
Well, actually, you can, for most auto parts There's no copyright in functional parts. For that, you need a patent. There is a big aftermarket auto parts industry. The parts are copies of the originals, not new designs. The major auto manufacturers have tried to get legislation to stop that in the US, but Congress rejected it.
I imagine we'll be seeing this used in autonomous vehicles pretty soon. Having four kinects on board with some basic processing could definitely replace some or all of the bulky LIDAR systems.
Won't work. The Kinect's structured light approach only works when you can overpower ambient light with the Kinect's little infrared source. That's indoor or night only, and short range. LIDAR units work in bright sunlight because, for a few nanoseconds per cycle, they're brighter than the sun in their portion of the spectrum.
Short range LIDAR units aren't bulky. They're smaller than the Kinect. For more range, you need bigger collecting optics.
Multiple-camera systems can now do stereo reasonably well without a structured illumination source. The purpose of the Kinect's dot pattern is to put some texture on uniform surfaces, like floors and walls, so the stereo algorithm can lock in. It doesn't matter what you project, as long as it has lots of edges in it. Outdoors, you usually have enough texture for stereo lockup. The most notable exception is fresh snow, for which humans sometimes have "white-out", losing depth perception.
The reactor may have melted through the base of its pressure vessel, but it's hard to tell. The high radiation levels could either be from a melt-through or from a leak as attempts are made to force water into the reactor pressure vessel. The latest JAIF status report contains almost all the hard data that's coming out. Everything else is secondary speculation based on that limited data.
No data seems to be available about pressure or temperature inside the reactor. That's listed as "unknown" for unit 2. The sensors involved were probably destroyed in one of the fires, explosions, or building collapses. Pressure in the containment vessel for unit 2 is listed as "low", whatever that means.
A full meltdown is now a real possibility. The JAIF chart has been showing "Fuel rods exposed partially or fully" for units 1, 2, and 3 for ten days now. Reactor pressure vessels are tough, as are containment structures, but ten days of no core cooling is well beyond design limits.
Understand that the water spraying operation refers to the containment structure, which is normally dry. Inside the containment is the reactor pressure vessel, which is a boiler. Getting water inside there, which is needed to cover the core and achieve cold shutdown, requires forcing it in against steam pressure. This has to be done in a highly radioactive environment, in a fire-ruined building where the walls and beams have collapsed, the pumps are damaged, and valves which are usually operated remotely have to be operated by people turning handwheels. Some people are trying very hard to do that. Some of them will probably die. If they succeed, there will be a local mess, but it will be manageable. If they fail, there will be a meltdown.
Google is smaller than it looks. The core search engine team was about 90-100 people as of a few years ago.
97% of the revenue still comes from search ads.
Google has a huge array of money-draining services, some of which are labor-intensive. They're not generating much revenue. Mostly, they're defensive measures to ward off Microsoft. GMail, Google Docs, the free hosting service, etc. exist to threaten Microsoft. It's not like offering spreadsheets on line is a viable business. Even the whole Android phone thing is mostly there to prevent Microsoft from monopolizing that space. (It's also a threat to Apple. Google pays Apple $100 million a year to stay on the iPhone. If it weren't for Android, Apple might provide their own closed iPhone search engine.)
Google spends an incredible amount of money on non-revenue defensive measures.
when the general public sees this kind of accident and some techie starts scoffing and arrogantly laughing and proclaiming how insignicant this accident is THEY STOP LISTENING TO YOU
The best reports on reactor status are at Japan Atomic Industrial Forum, which publishes a status table every day. This is addressed to people in the industry. They just list the facts, without explanation.
The good news for March 28 is that Unit 3's containment is now listed as "undamaged" instead of "possibly damaged". Unit 2 is listed as "damaged and leakage suspected", and that's now the most worrisome unit.
There's finally a fresh water supply for cooling. That's a big relief. Sea water cooling in a boil-off situation leaves tons of salt behind, and there was a real worry that the seawater cooling would stop working once too much salt accumulated. Fresh water cooling can continue indefinitely. It's not clear where the water is coming from. Hopefully they have a water line to a reliable source by now, and aren't just bringing in tanker trucks.
The cores in units 1,2, and 3 still have exposed fuel rods. Until water injection into the core is working again, the reactor can't be brought to cold shutdown. Remember, the reactor vessel is pressurized and contains a mixture of water and steam. Injecting water into a boiler is inherently difficult. Injecting water into a damaged boiler in a ruined structure in a highly radioactive area is very tough.
The spent fuel pool situation on reactors 3 and 4 is marginally under control. Seawater spray continues, but if they have to keep putting water in, the situation is still bad.
They're weeks from a stable emergency shutdown.
That's just the beginning. The situation isn't safe until there are again redundant closed loop cooling systems working. The current cooling hacks dump radioactive water into the ocean.
Then comes decommissioning. The spent fuel pools have to be cooled for three years or so, and then the fuel rods transferred from the wrecked buildings to dry casks. It will probably be necessary to build another containment building around unit 2, at least. Units 1,2, and 3 are all too damaged to ever de-fuel normally. It's not clear what will be done there.
Unit 4 wasn't fueled, but it had a hydrogen explosion while cooling was lost, and will probably never be restarted. Units 5 and 6 can potentially be restarted, but it's doubtful that they will be.
Festo does it again. Festo is a German robotics firm, and a very good one. Each year, they do a technical tour de force like this as a demo. They've previously done a swimming dolphin robot, a lighter-than-air flying jellyfish, and several other projects. Their overall direction is to learn to control very flexible systems, moving robotics away from the very rigid machines currently used.
American companies used to do things like that in the 1950s and 1960s. Japanese companies did that until the 1990s. When a country stops doing this, its technological dominance is over.
This guy doesn't need the "best programmers in the world". He's doing a feature-heavy payment card system that integrates with business expense reports. That's nice for people who travel too much, but it's not rocket science.
He's not doing cutting-edge technology, like machine learning, or autonomous robotics, or game engines for really big seamless shared worlds with intelligent NPCs, or modern high-end CAD. He needs competent people, but not people who are breaking new ground.
For what he's doing, ".NET" might not be a bad choice. They might have launched sooner.
Oh, that guy again. A few weeks ago, he was acting like an authority on web search. Last week, he was giving talks about how Boston is failing as a high-tech hub.
Firefox 4 is using 10% of 2 CPUs on Windows 7 when just sitting there displaying a completely static page. Firefox 3 went down to near 0 when idle. What's sucking up resources?
Requires an OS reboot, even on Windows 7, for installation. That's unnecessary, hasn't been necessary in years, and the Windows developer guidelines prohibit that.
Your choices are "no toolbar" or "really ugly four-line toolbar with mandatory tabbed browsing".
The instructions are a video.
More spam-friendly.
"Block images from this site" has disappeared as a right-click option.
This might potentially be a solution for spam. To send an email, you need a bitcoin. Bitcoins are easy to get in small quantities, maybe even free, but hard to get in bulk.
As a payment system, I don't see it. DigiCash had more promise as a distributed payment system, but Chaum blew the negotiations repeatedly.
There's something to be said for DC distribution within data center racks, but building a plug-in DC infrastructure seems like a PR stunt. They need a whole rack of power conversion gear to serve 50 desktop computers.
Google at one point proposed that rackmount computers should be built to run on 12VDC only, so you could have a single 12VDC supply in the rack and get rid of the individual power supplies for the server. Whatever happened to that?
Much industrial automation gear and military equipment runs off 24VDC. That's low enough that you don't have a shock hazard, but high enough that the wire sizes are reasonable.
Worse, it's an ad for a site which immediately wants you to sign up. I clicked "deny" for its cookie, and now the site won't load at all. Also, "an augmented browsing web app which allows you to see other people visiting the website you're visiting" sounds like a terrible idea from a privacy standpoint.
Google Voice has a pretty bad reputation for dealing with SMS messages,
That's been an ongoing problem. A year ago I dumped Google Voice for Twilo because of that. Google Voice gets its phone numbers from some third-tier telecom broker, one that has trouble identifying which numbers are considered "mobile" to the rest of the system. Getting their numbers via Sprint ought to help.
Mirah has semicolon trouble. It's an "you don't need semicolons, except where you do" language. Note that some of the import statements in Mirah end with semicolons. Either there's some obscure reason for that, or the code they're showing doesn't really work.
How do they do multiline statements? Please, not backslashes at the end of the line again. (Or worse, a syntax where a backslash followed by invisible whitespace, then a newline, has different semantics than without the whitespace.)
Python's indentation-based syntax seems to work out better, At least since the compiler got smart enough to understand when mixed spaces and tabs introduced visual ambiguity.
Mirah also seems to have a "you don't need declarations, except where you do" mindset. Historically, that's a bad decision. Almost every language that started out without declarations later backed into having them. FORTRAN, BASIC, C, and Perl all started out without much in the way of declarations. Python is one of the few typed but declaration-free languages that has succeeded. Arguably, Matlab is another.
Is Mirah really a front end to Java, or simply a dynamic language that targets the Java virtual machine? You can do something similar with.NET, running IronPython to the.NET VM.
Slashdot Mirror
Right. Especially since the Computer Museum in Sunnyvale already has a videogame section. They have the original PDP-1 Spacewar, the original Pong game, most of the early game consoles, etc. They even have some of the early games playable in emulators.
I thought Google's algorithms were supposed to decide which results were the best and put those at the top. I thought the "+1" was supposed to be a link to a site from another authoritative site. Is this an admission that this mechanism doesn't work? That it's been hopelessly gamed by spammers?
Basically, yes.
It's surprising that Google did this, after their disastrous experience with counting Yelp and CitySearch recommendations to compute ratings in Google Places. Last October, Google merged Places results from the previously independent Google Maps system into main web search results. Google Places ranking is based on location and recommendations, both of which can be easily faked and spammed. Within a month, we were seeing SEO firms advertising "Guaranteed first page listings or your money back". Within two months, Google Places spam was dominating results in local service categories like "dry cleaners" and "locksmiths".
Then Google was humiliated by the New York Times investigation into the J.C. Penny spam job. That was classic link farming, something Google supposedly knew how to stop. Google penalized J.C. Penny, but that created a political backlash. Now Google had admitted that they manually adjusted search results. The European Union's antitrust authorities started investigating that.
Blekko started sniping at Google for being so inept at dealing with web spam. Blekko improved results just by manually blocking the usual suspects, like "about.com". Demand Media went public and raised awareness of all the junk coming from "content farms". Google tried penalizing content farms, but it turned out that Google can't tell a content farm from a catalog shopping site, and was applying penalties to catalogs of real, buyable stuff. This angered some legitimate companies.
Google added a "mark as spam" link on search results a few weeks ago. (So they do have a "hate" button.) Now they've added a "like" button. They seem to be chasing after Blekko's "crowdsourced" search.
But it won't work. Blekko works only because nobody bothers to spam Blekko. They're too small to bother. Google's "like" mechanism will be spammed. Google can't stop people from creating massive numbers of phony Google accounts (see Jiffy Gmail Account Creator). As the previous Slashdot article points out, there are now far more social networking accounts than humans on the planet. Most of them were created by programs, not humans.
Really, though, Google has a "like" button because Facebook has a "like" button.
From the New Yorker article, here's the one big item of new info he's discovered:
In the standard historical accounts, the way that the bomb's gun mechanism worked was by shooting a cylindrical âoemaleâ uranium projectile into a concave, stationary uranium target. This act of atomic coitus created a mass sufficient to produce a critical reaction. The mass of the projectile was said to be 38.5 kilograms, and the mass of the target was said to be 25.6 kilograms. But no matter how many times Coster-Mullen did the math the numbers never quite worked out in a way that allowed the projectile and the target to fit inside the gun barrel while remaining subcritical.
The source of the error, Coster-Mullen recognized, was an assumption that every (male) researcher who studied the subject had made about the relation between projectile and target. These scholars had apparently been unable to conceive of an arrangement other than a "missionary position" bomb, in which a solid male projectile penetrated a vessel-like female target. But Coster-Mullen realized that a female-superior arrangement - in which a hollow projectile slammed down on top of a stationary cylinder of highly enriched uranium - yielded the correct size and mass.
Now that's a surprise. I wonder why the Manhattan Project did it that way, shooting the larger mass into the smaller mass. Maybe that was to get the assembly to hold together longer while the chain reaction initiated.
(For those of you who slept through the atomic weapons part of high school physics, the Hiroshima bomb was a "gun" bomb, where the tube from an artillery piece was used to fire one subcritical mass of uranium into another subcritical mass, producing a critical mass. That's been an obsolete technology for half a century, because such bombs are so bulky and require much more fissionable material than implosion bombs, but it works. The Trinity and Nagasaki bombs were implosion devices, of course.)
These frequent "big materials breakthrough" articles really should wait until they've been reviewed in some publication that knows something about the subject, like Chemical Engineering News. The paper, "Agro-waste nanocomposites for automotive applications", presented at the American Chemical Society is available. The claims there aren't as strong as the ones in the press release. Last year, the same author presented "Agro-Wastes Nanocomposites for Medical Application". Wonder what happened to that.
The trouble with many of these "new materials" is that they have some awful flaw. This one, for example, is "biodegradable". That means it rots. That's OK for packaging, but not for parts. Then there are basic questions, like will it tolerate water? Can it be made into thread, sheet, or film? Made at a reasonable cost?
There's been interest in finding useful things to make out of cellulose for the last century. There's so much agricultural waste around, and it would be nice to use it for something. Most of the ideas don't work out, but people keep trying.
France, Germany, and the US are all sending over rugged robots designed to work in disaster areas, collapsed building sites, and war zones. Strangely, Japan doesn't seem to have those.
I'm amazed that TEPCO hadn't at least brought in a few hobby-type R/C quadrotors or helicopters with TV cameras to get a look at areas they couldn't reach, like the spent fuel pools. For days, they didn't even know the water levels in those pools. Attempts were made to peer through holes in the roof with high-altitude flyovers, but that produced results like "we see some water in the pools, but can't tell how much."
The biggest problem with TEPCO has been eliminated. The CEO was just canned for "health reasons". He's been in hiding since the earthquake.
Just like you can't buy a car, reverse engineer it, and start mass-producing that car and competing with Mazda with their own product. You bought the car, not the plans to the car, nor the rights to sell it.
Well, actually, you can, for most auto parts There's no copyright in functional parts. For that, you need a patent. There is a big aftermarket auto parts industry. The parts are copies of the originals, not new designs. The major auto manufacturers have tried to get legislation to stop that in the US, but Congress rejected it.
I imagine we'll be seeing this used in autonomous vehicles pretty soon. Having four kinects on board with some basic processing could definitely replace some or all of the bulky LIDAR systems.
Won't work. The Kinect's structured light approach only works when you can overpower ambient light with the Kinect's little infrared source. That's indoor or night only, and short range. LIDAR units work in bright sunlight because, for a few nanoseconds per cycle, they're brighter than the sun in their portion of the spectrum.
Short range LIDAR units aren't bulky. They're smaller than the Kinect. For more range, you need bigger collecting optics.
Multiple-camera systems can now do stereo reasonably well without a structured illumination source. The purpose of the Kinect's dot pattern is to put some texture on uniform surfaces, like floors and walls, so the stereo algorithm can lock in. It doesn't matter what you project, as long as it has lots of edges in it. Outdoors, you usually have enough texture for stereo lockup. The most notable exception is fresh snow, for which humans sometimes have "white-out", losing depth perception.
And they probably don't know either.
The reactor may have melted through the base of its pressure vessel, but it's hard to tell. The high radiation levels could either be from a melt-through or from a leak as attempts are made to force water into the reactor pressure vessel. The latest JAIF status report contains almost all the hard data that's coming out. Everything else is secondary speculation based on that limited data.
No data seems to be available about pressure or temperature inside the reactor. That's listed as "unknown" for unit 2. The sensors involved were probably destroyed in one of the fires, explosions, or building collapses. Pressure in the containment vessel for unit 2 is listed as "low", whatever that means.
A full meltdown is now a real possibility. The JAIF chart has been showing "Fuel rods exposed partially or fully" for units 1, 2, and 3 for ten days now. Reactor pressure vessels are tough, as are containment structures, but ten days of no core cooling is well beyond design limits.
Understand that the water spraying operation refers to the containment structure, which is normally dry. Inside the containment is the reactor pressure vessel, which is a boiler. Getting water inside there, which is needed to cover the core and achieve cold shutdown, requires forcing it in against steam pressure. This has to be done in a highly radioactive environment, in a fire-ruined building where the walls and beams have collapsed, the pumps are damaged, and valves which are usually operated remotely have to be operated by people turning handwheels. Some people are trying very hard to do that. Some of them will probably die. If they succeed, there will be a local mess, but it will be manageable. If they fail, there will be a meltdown.
Google is smaller than it looks. The core search engine team was about 90-100 people as of a few years ago.
97% of the revenue still comes from search ads. Google has a huge array of money-draining services, some of which are labor-intensive. They're not generating much revenue. Mostly, they're defensive measures to ward off Microsoft. GMail, Google Docs, the free hosting service, etc. exist to threaten Microsoft. It's not like offering spreadsheets on line is a viable business. Even the whole Android phone thing is mostly there to prevent Microsoft from monopolizing that space. (It's also a threat to Apple. Google pays Apple $100 million a year to stay on the iPhone. If it weren't for Android, Apple might provide their own closed iPhone search engine.)
Google spends an incredible amount of money on non-revenue defensive measures.
when the general public sees this kind of accident and some techie starts scoffing and arrogantly laughing and proclaiming how insignicant this accident is THEY STOP LISTENING TO YOU
You mean like Jack Spencer, the Research Fellow in Nuclear Energy at The Heritage Foundation's Roe Institute for Economic Policy Studies?
Yes, this sounds like open source abandonware. That's not unusual; Google has done that a few times, too.
The best reports on reactor status are at Japan Atomic Industrial Forum, which publishes a status table every day. This is addressed to people in the industry. They just list the facts, without explanation.
The good news for March 28 is that Unit 3's containment is now listed as "undamaged" instead of "possibly damaged". Unit 2 is listed as "damaged and leakage suspected", and that's now the most worrisome unit.
There's finally a fresh water supply for cooling. That's a big relief. Sea water cooling in a boil-off situation leaves tons of salt behind, and there was a real worry that the seawater cooling would stop working once too much salt accumulated. Fresh water cooling can continue indefinitely. It's not clear where the water is coming from. Hopefully they have a water line to a reliable source by now, and aren't just bringing in tanker trucks.
The cores in units 1,2, and 3 still have exposed fuel rods. Until water injection into the core is working again, the reactor can't be brought to cold shutdown. Remember, the reactor vessel is pressurized and contains a mixture of water and steam. Injecting water into a boiler is inherently difficult. Injecting water into a damaged boiler in a ruined structure in a highly radioactive area is very tough.
The spent fuel pool situation on reactors 3 and 4 is marginally under control. Seawater spray continues, but if they have to keep putting water in, the situation is still bad.
They're weeks from a stable emergency shutdown.
That's just the beginning. The situation isn't safe until there are again redundant closed loop cooling systems working. The current cooling hacks dump radioactive water into the ocean.
Then comes decommissioning. The spent fuel pools have to be cooled for three years or so, and then the fuel rods transferred from the wrecked buildings to dry casks. It will probably be necessary to build another containment building around unit 2, at least. Units 1,2, and 3 are all too damaged to ever de-fuel normally. It's not clear what will be done there. Unit 4 wasn't fueled, but it had a hydrogen explosion while cooling was lost, and will probably never be restarted. Units 5 and 6 can potentially be restarted, but it's doubtful that they will be.
Festo does it again. Festo is a German robotics firm, and a very good one. Each year, they do a technical tour de force like this as a demo. They've previously done a swimming dolphin robot, a lighter-than-air flying jellyfish, and several other projects. Their overall direction is to learn to control very flexible systems, moving robotics away from the very rigid machines currently used.
American companies used to do things like that in the 1950s and 1960s. Japanese companies did that until the 1990s. When a country stops doing this, its technological dominance is over.
This guy doesn't need the "best programmers in the world". He's doing a feature-heavy payment card system that integrates with business expense reports. That's nice for people who travel too much, but it's not rocket science.
He's not doing cutting-edge technology, like machine learning, or autonomous robotics, or game engines for really big seamless shared worlds with intelligent NPCs, or modern high-end CAD. He needs competent people, but not people who are breaking new ground.
For what he's doing, ".NET" might not be a bad choice. They might have launched sooner.
Oh, that guy again. A few weeks ago, he was acting like an authority on web search. Last week, he was giving talks about how Boston is failing as a high-tech hub.
His actual career was in tools for modernizing COBOL programs. Even that wasn't too successful. Read "Mouth piece: Vivek Wadhwa''s talent for trumpeting his company shines, but observers want to see another kind of performance ".
If you're on the phone, or playing a game, you don't need music. (Or cigarettes. Phone usage has made a big dent in young people smoking.) Music competes with Farmville and Angry Birds now.
Firefox 4 is using 10% of 2 CPUs on Windows 7 when just sitting there displaying a completely static page. Firefox 3 went down to near 0 when idle. What's sucking up resources?
Not too impressed with Firefox 4.
This might potentially be a solution for spam. To send an email, you need a bitcoin. Bitcoins are easy to get in small quantities, maybe even free, but hard to get in bulk.
As a payment system, I don't see it. DigiCash had more promise as a distributed payment system, but Chaum blew the negotiations repeatedly.
There's something to be said for DC distribution within data center racks, but building a plug-in DC infrastructure seems like a PR stunt. They need a whole rack of power conversion gear to serve 50 desktop computers.
Google at one point proposed that rackmount computers should be built to run on 12VDC only, so you could have a single 12VDC supply in the rack and get rid of the individual power supplies for the server. Whatever happened to that?
Much industrial automation gear and military equipment runs off 24VDC. That's low enough that you don't have a shock hazard, but high enough that the wire sizes are reasonable.
Worse, it's an ad for a site which immediately wants you to sign up. I clicked "deny" for its cookie, and now the site won't load at all. Also, "an augmented browsing web app which allows you to see other people visiting the website you're visiting" sounds like a terrible idea from a privacy standpoint.
self generated certs
Man in the middle attack.
"A self-generated cert lets you talk encrypted to your attacker."
Google Voice has a pretty bad reputation for dealing with SMS messages,
That's been an ongoing problem. A year ago I dumped Google Voice for Twilo because of that. Google Voice gets its phone numbers from some third-tier telecom broker, one that has trouble identifying which numbers are considered "mobile" to the rest of the system. Getting their numbers via Sprint ought to help.
Mirah has semicolon trouble. It's an "you don't need semicolons, except where you do" language. Note that some of the import statements in Mirah end with semicolons. Either there's some obscure reason for that, or the code they're showing doesn't really work. How do they do multiline statements? Please, not backslashes at the end of the line again. (Or worse, a syntax where a backslash followed by invisible whitespace, then a newline, has different semantics than without the whitespace.)
Python's indentation-based syntax seems to work out better, At least since the compiler got smart enough to understand when mixed spaces and tabs introduced visual ambiguity.
Mirah also seems to have a "you don't need declarations, except where you do" mindset. Historically, that's a bad decision. Almost every language that started out without declarations later backed into having them. FORTRAN, BASIC, C, and Perl all started out without much in the way of declarations. Python is one of the few typed but declaration-free languages that has succeeded. Arguably, Matlab is another.
Is Mirah really a front end to Java, or simply a dynamic language that targets the Java virtual machine? You can do something similar with .NET, running IronPython to the .NET VM.
Do real experiments. The kids will remember that.