If you bought into any of these, you're a sucker. They don't work any more.
Divx (1998-2001). "Disposable" DVDs tied to a remote authorization system. Promoted by Circuit City and Thompson. Content now unplayable.
WalMart Music (2007-2008) Downloadable music tied to an authorization server. Content now unplayable.
PlaysForSure (2004-2008) Microsoft system. Downloadable music tied to an authorization server. Content from AOL MusicNow (closed), Musicmatch Jukebox (closed), Yahoo! Music Unlimited (closed), Spiralfrog (closed), MTV URGE (closed), MSN Music (closed), Musicmatch Jukebox (closed), Ruckus Network (closed) now generally unplayable, although exit strategies exist. Authorization servers were to be shut down August 31, 2008, but were kept up after that date.
General Motors tried it with Firebird III in 1958. That car also featured automatic lane-holding, using a wire in the pavement. The vehicle was very hard to control, but it looked really cool.
When we did our DARPA Grand Challenge vehicle, at one point we were controlling it remotely over a WiFi link for test purposes. We tried a joystick, but it was too easy to overcontrol. We got a Logitech racing game steering wheel (a USB peripheral) and pedals, interfaced those, and the thing was easily driveable remotely. That was without force feedback on the wheel, incidentally. (That wheel has "force feedback", but it's set up to take an audio track for vibration, not a desired position, so the HID interface doesn't properly support positional feedback.)
mcgrew.info blocked by current "robots.txt" file. The Archive treats "robots.txt" files as retroactive; if the current "robots.txt" won't allow archiving, then the Archive won't display old archived copies. The data is still in the Archive, but not publicly visible.
The entire content of the Internet fits in a 20x8x8 box operated by the Internet Archive. Cuil, which searches as much of the Web as Google, has one relatively modest data center. About half the system does the crawl and builds the index; the other half answers queries. So Google's main search engine function doesn't really require that much capacity by current standards. Of course, Google has a huge number of query servers front-ending the main index, which is of course replicated.
Why does Google need so much server capacity? YouTube? Command completion? GMail spam filtering? Ad serving?
Some sounds really suck at low bit rates, like cymbals. All the info is in the higher frequencies.
One of the worst cases is voice over white noise. Humans are good at pulling voice out of white noise, but most codecs use too many bits trying to replicate the white noise.
If this had been done by Hayao Miyazaki, the director of "Spirited Away", it might have been
good. He does kids as lead characters very well. But Miyazaki doesn't need to do remakes. He can develop original concepts.
The director of "Flushed Away"? Much lower down the food chain.
If you think software developers are weird, you're not getting out enough.
Commission salespeople and futures traders are much weirder. Some CEOs are weird. Low-end rock musicians are weird. (Above the "club band" level, some sanity tends to emerge, or at least the self-destructive ones are filtered out.) Strippers are weird. Successful high-end call girls, though, tend to be chillingly sane when not in their work personas.
That's almost as good as the Bladestar, available now for only $24.95 Yes, an indoor helicopter with radio control and collision avoidance for $24.95. WowWee manages to produce incredible capabilities at very low price points.
The MIT thing is neat, but it's mostly possible because IMU units and laser scanners are finally small enough for this.
The time-of-flight laser scanner thing has been frustrating. That's what you want in a robot; they're just too expensive for volume or hobbyist use. 3D LIDAR scanners have been around since the 1980s. For a long time, the SICK LMS devices, which worked well but were both bulky and expensive, ($9000), dominated the field. The DARPA Grand Challenge resulted in the Velodyne scanner, which was 3D (the SICK units are line scanners), but that cost $50,000 - $100,000.
There's no fundamental reason why the things should cost that much, except that they are produced in small volumes.
Back in 2004, I dragged a venture capitalist down to Advanced Scientific Concepts, which has a $100,000 flash laser rangefinder with a custom imaging chip. That thing should cost under $1000 in quantity, and eventually it should cost like a webcam. But ASC was selling only to the military, and they weren't thinking in terms of a volume product. It was too early, though; no volume market was on the horizon.
When robots at the Roomba level get laser rangefinding and a decent IMU, they'll be able to navigate without getting lost or stuck. An automatic vacuum with enough smarts to do the job right could take over much commercial floor care.
There are three issues with "online anonymity". One is anonymous businesses, the second is the ability to create an unlimited number of new identities at very low cost, and the third is actual identification of end users.
Anonymous businesses, that is, web sites with commercial intent which don't identify their ownership, are already illegal in manyjurisdictions. At SiteTruth, we treat anonymous businesses (where there's no postal mailing address on the web site) as "bottom feeders", and move them to the bottom of search results. Google has a bias against "private registration" domains, but that only kicks in if the site otherwise looks like a junk site. There's not much controversy about this; it's accepted law that a business has to identify itself properly.
The ability to create an unlimited number of new identities causes various forms of trouble. The ability to get vast numbers of free Gmail accounts ("automatically create Gmail Accounts in seconds flat without breaking a sweat") is a windfall for spammers and has destroyed vast sections of Craigslist. The ability to register large numbers of domains with phony domain registration has created a well-known range of problems. Gradually, that's being tightened down. "Domain Tasting" is now dead, now that registrars have to eat the loss if they register and release a domain within 5 days. Phony WHOIS information remains a problem, but could be fixed. When you register a domain, you should get a postal mail piece with the code that enables the domain.
End user identification is the controversial issue. The music industry would like it, but, after all, the music industry is a dinky business compared to the Internet. IBM, HP, Dell, Microsoft, Yahoo, and Google are each bigger than the entire music industry. Other than for email sending, there's other big interest behind end user identification.
Actually, the whole system is getting close to 1.8K, but some magnets aren't quite down there yet. About 2/3 of the ring has cyro authorization (cold enough to power up the magnets) but the magnets haven't been energized yet. All the magnets have to be powered up. Then comes low power beam testing and alignment. Then maybe they can do some science.
There are supposed to be two big fixes in place now. First, the quench protection system now covers not just the magnets, but the connections to them. (The basic idea is that if a superconducting magnet ceases to be superconductive at some hot spot (in which case all the energy in the magnet comes out as heat), the system dumps the energy into resistive loads, and heats up the entire magnet quickly to make it resistive, so that the energy is dumped throughout the magnet, not just at the hot spot. Last time, a hot spot developed at a welded splice. Second, the venting system for dealing with the gaseous helium released after a quench has been improved, with bigger rupture discs. Last time, the vents weren't big enough, and there was substantial damage to the cryogenic plumbing.
None of this has anything to do with the physics. It's all plumbing and DC power control.
The original design documents say a quench is supposed to be recoverable within three hours. That was rather optimistic.
Google SecurityView - security cameras can be fed into Google, and will appear as live inserts. Total coverage of all 7-11 and Starbucks outlets.
Google AndroidView - all Google Android phones will take pictures of their surroundings when not otherwise busy, to be loaded into StreetView.
Google HiRes - updates to StreetView will all be with the newer high resolution cameras, replacing the low-rez images still up for many older locations.
I'm going to get flamed by the usual losers for this, but anyway...
I hold three software patents. One was sold as part of a deal that made me several million dollars back in the 1980s. One made me $600,000 in licensing fees. I'm pursuing an infringement claim against DoD on the third.
I also have another patent pending. I put "Inventor" on my tax return.
Each of these patents was an early patent in an area where previous attempts to solve the problem had failed. In each case, the patented technology came with a working application or a successful demo. So these weren't just "ideas", they were ideas that worked. That's when patents are worthwhile - you've solved a hard problem, you're not with a big enough company to exploit it, and doing a startup doesn't seem to be the right answer.
The key point here is that a patent plus a demo version is more valuable than either alone.
This history isn't all that unusual here in Silicon Valley.
Google finally fixed this. The offending page now reads "We're sorry. You can't access this spreadsheet because it is in violation of our Terms of service. If you feel this is in error, please contact us."
Sometimes you just have to use a big clue stick to get their attention. It took some help from The Register to get Yahoo, Microsoft, and eBay to clean up their acts.
Five more long-term exploited sites remain. A bit more nagging, and we'll have this cleaned up.
Once this is cleaned up, phishing blacklists that blacklist entire second-level domains will be effective. No more just blacklisting the URL.
The trouble with supercomputing is that, if you have to share the thing, you don't need it.
Supercomputers are worth the trouble if there are applications that need hours or days of time. But if
you have many users sharing the thing, it's a waste. Price/performance tends to be maximized towards
the upper end of mainstream machines. Supercomputers, with their custom hardware, tend to have lower
price/performance than commodity machines. That's why web farms are made of commodity hardware.
Google has a malware hosting problem of their own.
Google Spreadsheets can be abused to create phony login pages. Here's one for "Free Habbo credits", designed to collect Habbo logins.
It's been reported via the usual "Google abuse" mechanism, repeatedly, and it's still up. It's been up since
October 28, 2008.
We track major domains being exploited by active phishing scams. ("Major" here means only that it's in Open Directory, with about 1.5 million domains.) There are 39 exploited domains today. Only 7 have been on that list since 2008. The most abused site is Piczo.com, which is a hosting service/social network/shopping site for teenagers.
Just about everybody else has cleaned up their act. 18 months ago, that list had 174 entries, including Yahoo, eBay, Microsoft Live, and TinyURL. All those companies have become more aggressive about checking for phishing scams that were injected into their domain. Google's cluelessness in this area ought to be embarrassing to someone.
He's so right. He references the Turkey City Lexicon, which lists most of the things that make bad SF. Also worth reading is the Evil Overlord List. (" 2. My ventilation ducts will be too small to crawl through." "56. My Legions of Terror will be trained in basic marksmanship. Any who cannot learn to hit a man-sized target at 10 meters will be used for target practice." "67. No matter how many shorts we have in the system, my guards will be instructed to treat every surveillance camera malfunction as a full-scale emergency.")
There are some other annoying cliches in SF. One is copying historical battles. The Defense of Roarke's Drift has shown up in at least four SF novels. (Nobody ever seems to do the Defense of Duffer's Drift.) Star Wars space battles are copied from WWI biplane battles, where nobody can hit targets consistently, even at short range. There's also the embarrassing fact that, historically, heroism hasn't decided many major battles. (Roman saying: "The Legion is not composed of heroes. Heroes are what the Legion kills.") Military SF no longer reflects this, because the WWII generation, which learned that the hard way, has died off.
David Weber does battles better, but his stuff requires too much exposition for most people. His latest book in the Honor Harrington series consists mostly of transcripts of meetings, setting up the political background for the next book.
Stross himself has his moments. The Merchant's War series starts out as fantasy, but slowly, book by book, moves into hard fiction and then politics. In the last book out so far, a character modelled on Dick Cheney has dealt with a threat from a castle in an alternate universe by having his people blow up the castle with a nuclear weapon.
If you work for a regulated gambling operation in the developed world, you'll probably have to undergo a background check. That's a plus. On the other hand, working for some "offshore online casino" is iffy, and may even result in painful interviews with law enforcement and tax authorities.
One of the interesting aspects of gambling software is that many of the companies that provide services are held financially responsible for both errors and downtime. It's not like the rest of the on-line world, where everything is the customer's fault. If you want to see high-reliability systems, check out the data center behind one of the big state Lotto operations.
One problem is that CA's expect "instant gratification" when they issue a new root certificate.
Properly, root certificates ought to change very rarely. They should be published a few years before they're used, so there's time for people to install new browser updates and be confident that a fake root cert didn't get through. Now, though, Microsoft Windows Update is being used as the source of root CA authority. We're seeing far more root certificates, more than are really necessary.
The U.S. Department of Defense is going to all-HTTPS for most military sites, even the public-facing ones. But they have far too many root certificates, and few browsers have them all.
Peer-weighted reputation, like "web of trust" systems, won't work if the "peers" are anonymous. Otherwise, we get link farms and similar forms of bulk spamming.
Even without anonymity, imagine Rush's "dittoheads" as a source of authority for news.
If they'd given out, say, 10 copies of Windows 7, 10 copies of a game that uses Windows 7 in some useful way, and a 10-port LAN hub, at least you could have a LAN party.
Then again, you can do it right. GTA not only has ads and commercials for fake companies, it has the businesses behind them. You can get a Whiz Wireless cell phone in-game. You can go to Burger Shot. That's doing it right.
Slashdot Mirror
If you bought into any of these, you're a sucker. They don't work any more.
Next, Disney.
Joystick steering has been tried, and it sucks.
General Motors tried it with Firebird III in 1958. That car also featured automatic lane-holding, using a wire in the pavement. The vehicle was very hard to control, but it looked really cool.
When we did our DARPA Grand Challenge vehicle, at one point we were controlling it remotely over a WiFi link for test purposes. We tried a joystick, but it was too easy to overcontrol. We got a Logitech racing game steering wheel (a USB peripheral) and pedals, interfaced those, and the thing was easily driveable remotely. That was without force feedback on the wheel, incidentally. (That wheel has "force feedback", but it's set up to take an audio track for vibration, not a desired position, so the HID interface doesn't properly support positional feedback.)
I'd hazard a guess you won't find mcgrew.info or holy-bible.us there, either.
The entire content of the Internet fits in a 20x8x8 box operated by the Internet Archive. Cuil, which searches as much of the Web as Google, has one relatively modest data center. About half the system does the crawl and builds the index; the other half answers queries. So Google's main search engine function doesn't really require that much capacity by current standards. Of course, Google has a huge number of query servers front-ending the main index, which is of course replicated.
Why does Google need so much server capacity? YouTube? Command completion? GMail spam filtering? Ad serving?
They need to try more samples.
Some sounds really suck at low bit rates, like cymbals. All the info is in the higher frequencies.
One of the worst cases is voice over white noise. Humans are good at pulling voice out of white noise, but most codecs use too many bits trying to replicate the white noise.
If this had been done by Hayao Miyazaki, the director of "Spirited Away", it might have been good. He does kids as lead characters very well. But Miyazaki doesn't need to do remakes. He can develop original concepts.
The director of "Flushed Away"? Much lower down the food chain.
For a good cartoon remake, see Tex Avery's Red Hot Riding Hood.
You know lots of CEOs, futures traders, rock musicians, and high-end call girls, do you? I'd like to work where you work.
I was in San Francisco for the dot-com boom.
If you think software developers are weird, you're not getting out enough.
Commission salespeople and futures traders are much weirder. Some CEOs are weird. Low-end rock musicians are weird. (Above the "club band" level, some sanity tends to emerge, or at least the self-destructive ones are filtered out.) Strippers are weird. Successful high-end call girls, though, tend to be chillingly sane when not in their work personas.
That's almost as good as the Bladestar, available now for only $24.95 Yes, an indoor helicopter with radio control and collision avoidance for $24.95. WowWee manages to produce incredible capabilities at very low price points.
The MIT thing is neat, but it's mostly possible because IMU units and laser scanners are finally small enough for this.
The time-of-flight laser scanner thing has been frustrating. That's what you want in a robot; they're just too expensive for volume or hobbyist use. 3D LIDAR scanners have been around since the 1980s. For a long time, the SICK LMS devices, which worked well but were both bulky and expensive, ($9000), dominated the field. The DARPA Grand Challenge resulted in the Velodyne scanner, which was 3D (the SICK units are line scanners), but that cost $50,000 - $100,000.
There's no fundamental reason why the things should cost that much, except that they are produced in small volumes. Back in 2004, I dragged a venture capitalist down to Advanced Scientific Concepts, which has a $100,000 flash laser rangefinder with a custom imaging chip. That thing should cost under $1000 in quantity, and eventually it should cost like a webcam. But ASC was selling only to the military, and they weren't thinking in terms of a volume product. It was too early, though; no volume market was on the horizon.
When robots at the Roomba level get laser rangefinding and a decent IMU, they'll be able to navigate without getting lost or stuck. An automatic vacuum with enough smarts to do the job right could take over much commercial floor care.
There are three issues with "online anonymity". One is anonymous businesses, the second is the ability to create an unlimited number of new identities at very low cost, and the third is actual identification of end users.
Anonymous businesses, that is, web sites with commercial intent which don't identify their ownership, are already illegal in many jurisdictions. At SiteTruth, we treat anonymous businesses (where there's no postal mailing address on the web site) as "bottom feeders", and move them to the bottom of search results. Google has a bias against "private registration" domains, but that only kicks in if the site otherwise looks like a junk site. There's not much controversy about this; it's accepted law that a business has to identify itself properly.
The ability to create an unlimited number of new identities causes various forms of trouble. The ability to get vast numbers of free Gmail accounts ("automatically create Gmail Accounts in seconds flat without breaking a sweat") is a windfall for spammers and has destroyed vast sections of Craigslist. The ability to register large numbers of domains with phony domain registration has created a well-known range of problems. Gradually, that's being tightened down. "Domain Tasting" is now dead, now that registrars have to eat the loss if they register and release a domain within 5 days. Phony WHOIS information remains a problem, but could be fixed. When you register a domain, you should get a postal mail piece with the code that enables the domain.
End user identification is the controversial issue. The music industry would like it, but, after all, the music industry is a dinky business compared to the Internet. IBM, HP, Dell, Microsoft, Yahoo, and Google are each bigger than the entire music industry. Other than for email sending, there's other big interest behind end user identification.
Actually, the whole system is getting close to 1.8K, but some magnets aren't quite down there yet. About 2/3 of the ring has cyro authorization (cold enough to power up the magnets) but the magnets haven't been energized yet. All the magnets have to be powered up. Then comes low power beam testing and alignment. Then maybe they can do some science.
There are supposed to be two big fixes in place now. First, the quench protection system now covers not just the magnets, but the connections to them. (The basic idea is that if a superconducting magnet ceases to be superconductive at some hot spot (in which case all the energy in the magnet comes out as heat), the system dumps the energy into resistive loads, and heats up the entire magnet quickly to make it resistive, so that the energy is dumped throughout the magnet, not just at the hot spot. Last time, a hot spot developed at a welded splice. Second, the venting system for dealing with the gaseous helium released after a quench has been improved, with bigger rupture discs. Last time, the vents weren't big enough, and there was substantial damage to the cryogenic plumbing.
None of this has anything to do with the physics. It's all plumbing and DC power control.
The original design documents say a quench is supposed to be recoverable within three hours. That was rather optimistic.
This is just the beginning. Next up,
I'm going to get flamed by the usual losers for this, but anyway...
I hold three software patents. One was sold as part of a deal that made me several million dollars back in the 1980s. One made me $600,000 in licensing fees. I'm pursuing an infringement claim against DoD on the third. I also have another patent pending. I put "Inventor" on my tax return.
Each of these patents was an early patent in an area where previous attempts to solve the problem had failed. In each case, the patented technology came with a working application or a successful demo. So these weren't just "ideas", they were ideas that worked. That's when patents are worthwhile - you've solved a hard problem, you're not with a big enough company to exploit it, and doing a startup doesn't seem to be the right answer.
The key point here is that a patent plus a demo version is more valuable than either alone.
This history isn't all that unusual here in Silicon Valley.
Google finally fixed this. The offending page now reads "We're sorry. You can't access this spreadsheet because it is in violation of our Terms of service. If you feel this is in error, please contact us."
Sometimes you just have to use a big clue stick to get their attention. It took some help from The Register to get Yahoo, Microsoft, and eBay to clean up their acts.
Five more long-term exploited sites remain. A bit more nagging, and we'll have this cleaned up.
Once this is cleaned up, phishing blacklists that blacklist entire second-level domains will be effective. No more just blacklisting the URL.
The trouble with supercomputing is that, if you have to share the thing, you don't need it.
Supercomputers are worth the trouble if there are applications that need hours or days of time. But if you have many users sharing the thing, it's a waste. Price/performance tends to be maximized towards the upper end of mainstream machines. Supercomputers, with their custom hardware, tend to have lower price/performance than commodity machines. That's why web farms are made of commodity hardware.
An ordinary scam (like the Habbo one listed above) is different from a phishing attack (which requires that the attacker impersonates another entity).
PhishTank calls it a phishing scam. We follow their data.
Google has a malware hosting problem of their own.
Google Spreadsheets can be abused to create phony login pages. Here's one for "Free Habbo credits", designed to collect Habbo logins. It's been reported via the usual "Google abuse" mechanism, repeatedly, and it's still up. It's been up since October 28, 2008.
We track major domains being exploited by active phishing scams. ("Major" here means only that it's in Open Directory, with about 1.5 million domains.) There are 39 exploited domains today. Only 7 have been on that list since 2008. The most abused site is Piczo.com, which is a hosting service/social network/shopping site for teenagers.
Just about everybody else has cleaned up their act. 18 months ago, that list had 174 entries, including Yahoo, eBay, Microsoft Live, and TinyURL. All those companies have become more aggressive about checking for phishing scams that were injected into their domain. Google's cluelessness in this area ought to be embarrassing to someone.
He's so right. He references the Turkey City Lexicon, which lists most of the things that make bad SF. Also worth reading is the Evil Overlord List. (" 2. My ventilation ducts will be too small to crawl through." "56. My Legions of Terror will be trained in basic marksmanship. Any who cannot learn to hit a man-sized target at 10 meters will be used for target practice." "67. No matter how many shorts we have in the system, my guards will be instructed to treat every surveillance camera malfunction as a full-scale emergency.")
There are some other annoying cliches in SF. One is copying historical battles. The Defense of Roarke's Drift has shown up in at least four SF novels. (Nobody ever seems to do the Defense of Duffer's Drift.) Star Wars space battles are copied from WWI biplane battles, where nobody can hit targets consistently, even at short range. There's also the embarrassing fact that, historically, heroism hasn't decided many major battles. (Roman saying: "The Legion is not composed of heroes. Heroes are what the Legion kills.") Military SF no longer reflects this, because the WWII generation, which learned that the hard way, has died off.
David Weber does battles better, but his stuff requires too much exposition for most people. His latest book in the Honor Harrington series consists mostly of transcripts of meetings, setting up the political background for the next book.
Stross himself has his moments. The Merchant's War series starts out as fantasy, but slowly, book by book, moves into hard fiction and then politics. In the last book out so far, a character modelled on Dick Cheney has dealt with a threat from a castle in an alternate universe by having his people blow up the castle with a nuclear weapon.
Mod parent up.
If you work for a regulated gambling operation in the developed world, you'll probably have to undergo a background check. That's a plus. On the other hand, working for some "offshore online casino" is iffy, and may even result in painful interviews with law enforcement and tax authorities.
One of the interesting aspects of gambling software is that many of the companies that provide services are held financially responsible for both errors and downtime. It's not like the rest of the on-line world, where everything is the customer's fault. If you want to see high-reliability systems, check out the data center behind one of the big state Lotto operations.
One problem is that CA's expect "instant gratification" when they issue a new root certificate. Properly, root certificates ought to change very rarely. They should be published a few years before they're used, so there's time for people to install new browser updates and be confident that a fake root cert didn't get through. Now, though, Microsoft Windows Update is being used as the source of root CA authority. We're seeing far more root certificates, more than are really necessary.
The U.S. Department of Defense is going to all-HTTPS for most military sites, even the public-facing ones. But they have far too many root certificates, and few browsers have them all.
peer-weighted reputation...
Peer-weighted reputation, like "web of trust" systems, won't work if the "peers" are anonymous. Otherwise, we get link farms and similar forms of bulk spamming.
Even without anonymity, imagine Rush's "dittoheads" as a source of authority for news.
If they'd given out, say, 10 copies of Windows 7, 10 copies of a game that uses Windows 7 in some useful way, and a 10-port LAN hub, at least you could have a LAN party.
Then again, you can do it right. GTA not only has ads and commercials for fake companies, it has the businesses behind them. You can get a Whiz Wireless cell phone in-game. You can go to Burger Shot. That's doing it right.
The print version of that article has before and after pictures.
Where? I don't see any.
Only the actual paper version has the pictures.