MIT runs a class called 6.033: Computer Systems Engineering. These lecture notes contain a list of projects that had great sums of money spent on them only to be abandoned. Also the reading list has a bunch of papers that discuss the "big splash" failures like Therac 25.
This has got to be a joke! Listen, I hate to be insulting, but the odds are about 1:1,000,000 to one that the breakthrough you think you have is nothing of the sort. It's true I know close to nothing about you, but the name you chose (which implies re-use of one time pads), the question you ask and the fact that Ask Slashdot seems like an appropriate forum tells me that you're an amateur.
Go read back issues of Crypto-Gram and read up on all the lame hype laden "unbreakable" crypto schemes (often based on one time pads) that they destroy and then laugh at. If after reading all that you're still convinced you've got something, sure, go see a patent attorney.
My guess is you'll end up saving yourself the patent fees and a fair amount of humiliation by just letting it drop.
Quoting from the article on the GM site: "All of AUTOnomy's essential systems, including the fuel cell stack and on-board hydrogen storage system, are neatly packaged in the skateboard-like chassis."
This is only slightly less optimistic than talking about installing a Mr. Fusion. The "on-board hydrogen storage system" (that fits in a six inch frame, no less!) is currently total fantasy. So there are a few things left to work out before we all get to drive cars like this.
Well hold on a second there... I wouldn't say the existing system worked just fine. Even with paper, there's good design and bad design. Those butterfuly ballots were confusing as hell and certainly needed to be redesigned.
Also needing to be addressed were the charges of systematic voter disenfranchisement, the polling places that closed early and so on.. but these are issues regardless of the implementation technology for the voting itself.
I'm a big fan of technology, make my living at it, love linux, etc... but I've never been convinced that electronic voting is in any way superior to old-fashioned voting.
Let me describe the voting system Canada has: You register much as you do here. You show up at the polling place. They cross your name off the list and hand you a hard to forge ballot. You walk behind a little screen, put an X next to the person you want to vote for and stick it in a box. At the end of the day, representatives from each party and the media open the box and count the ballots. The results are delivered in a tree - local place reports to city, city probably to county, county to province. They add up all the results and they declare a winner.
Nothing about this fails to scale. In other words, a population 10x the size of Canada requires about 10x the number of volunteers which works out to be the same number of volunteers per capita.
This system seems so much more workable to me, there are so many fewer opportunities for breakdown.
Is it Auditable? Yes, keep the ballots locked up and recount them.
Is it anonymous? Yes, at least as much as touch screen voting.
Is there any software / printers / touchscreens / whatever to fail? No.
Why do we need millions of dollars of development and plenty of technology to fail when a bunch of pieces of paper and some pens would do fine?
I'd recommend you check out www.pcsforeveryone.com which is in the Cambridge MA area. They provide "near custom" services where you can put together pcs at almost the level of detail you would when doing it yourself (Like choosing which fan, which case, etc...) They're pretty clueful, know about Linux and can pre-install it, won't charge you for an OS if you don't want one, etc...
Even if you don't buy from them, it's worth reading their commentary on which parts are best and which parts work together / not with linux / etc...
Yes, it's using Unary, but you want to be able to compute "cost per byte" right? So somehow we need to be able to compare the value of this message to other binary encoded messages like the password for an Everquest account. This requires your message to encoded in binary.
However, if you want to be super picky, you can figure out the fractional requirement for how many bits are required to encode this message with this formula:
log (Number of states) / log (Number of states per symbol)
Which in this case works out to log 3 / log 2
or approximately 1.58 bits. Of course if this is the only message you're sending (which it was) you'd need to round up to the next integral number which brings us back to two.
From the ride of Paul Revere. So if you consider this three possible states (0-2 lanterns) it takes 2 bits to encode.
What was the "value" of that particular message? Up for debate of course but conventional wisdom is that the revolutionary war would have been lost without it and history would be very different.
The thing about Music is that nothing about it is _inherently_ expensive to produce. Sure, once you've thrown in the videos and the launch party and the services of the London Philharmonic it starts to get up there (Although I'd guess the services of the London Philharmonic are cheaper than you'd think) but the equipment, space and talent to just record music is generally within the range of every day people. Of course certain kinds of music are easier to record on the cheap than others (Moby can do it alone in his apartment because, well, he doesn't have any instruments) but with a nice Mac and $10K worth of extra hardware a talented bunch of people can put out some pretty respectable stuff. So music will live on, even if it's almost free. Same goes for books.
But what about Movies? Movies are going to be subject to the very same dynamic, although perhaps timeshifted a few years to the right. If the shit start to hit the movie industry, the world is going to start to look pretty different because movies are _expensive_. I mean even once you throw out the union pay scales and the staffing and the rules, blowing shit up (which is a staple of a lot of movies) is expensive; As are sets and crowds and animations and all the other stuff we see in our movies. Sure, you can still make "Clerks" and "The Blair Witch Project" pretty cheaply, but those aren't the only kinds of movies out there - not even an appreciable percentage if you're looking at Hollywood output. So the big budget movie could be a thing of the (soon to be) past.
Maybe movies will go all digital. Computing cycles will be so cheap and software so good that movies can be "filmed" at low cost by some Savant in his basement with the futuristic equivalent of an iMac and some Red Bull. But I wouldn't count on it.
It's a little more complicated than that. Christensen's contribution was that incumbent players miss disruptive technologies not because of bad management, but because of good management. The disruptive technologies initially are inferior, serve a smaller market and come with low margins and immature technology. At no point is it a rational decision to dump your successful high margin technology for an inferior product with a lower margin and a smaller market.
But what happens is that the disruptive technology improves with time to the point that it's "good enough" for your customers, even though it still can't do all the things the incumbent technology can do. And presto, the market disappears overnight.
I think a good example (although it's largely conjecture at this point) is the storage industry. EMC used to make a fortune selling extremely fast, reliable, high margin storage devices to enterprises. There have always been commodity PCs that did fileserving, but it was a different market: Less space, less reliable, less manageable, more failure prone! Who cares if it's cheaper.
Well over time, commodity PCs improve, technology gets better and people start to think, why not just have lots of commodity PCs and some clever software engineering stand in for my expensive EMC. Google launches and proves (to the people paying attention) that clusters of commodity PCs with clever software can be pretty damn fast and pretty reliable at storing rather a lot of data. And the cost is phenomenal.
At this point I'd say the writing is on the wall, but is EMC building commodity based storage solutions? Maybe in some tiny little skunkworks project, but not visibly. Even if they build it, will they launch it? Not clear. Why would they? They sell a Symmetrix for >$1M, the margins are huge on those things. And PCs are never going to be as fast as a tuned Symmetrix and fibre channel connection.
P2P research by well known research institutions is far from unheard of. MIT has Chord which is a project to produce robust scalable distributed systems using peer to peer ideas. They have an efficient hash lookup algorithm that could form the basis of many p2p systems and they have code available for download.
The great thing (for us) about this argument is that it's really hard to prove what the market effect will be.
Of course they'll argue what they think the market effect will be, but then there's the simple question, "Didn't you guys say the same thing about the VCR?"
Actually they just launched a new version a couple of weeks ago - it's pretty cool and they seem to do a much better job of picking songs. It lists what all your friends are listening to, so if you have a couple of friends listening, someone is always listening to something cool and you can just jump right in with them.
This is a horrible slippery slope and shame on you for not recognizing it.
1) An.xxx tld is set up so that porn companies can disitnguish themselves if they so choose.
2) AOL, libraries, schools and a bunch of others filter.xxx domains at the router.
3) Porn companies start getting pressure that they must register themselves in the.xxx domain. Eventually a law is passed.
4) Someone somewhere is tasked with deciding what qualifies under the mandatory.xxx law and what doesn't. While not being censored per se anything in the.xxx domain is totally marginalized because nobody can see it.
Is Mapplethorpe Porn? Is AIDS awareness? Is birth control information? What is porn? Let me guess, you'll know it when you see it right?
I think a real key here is that different students have vastly differing skill levels, so giving an assignment that can be handled meaningfully at all levels is great.
We did this when I was in highschool, and the most interesting (and gratifying) thing was the social impact. Video games were banned from our (and many other) High School computer labs, but games written by students were allowed.
I wrote a 4 player version of light cycles of tron, and it was pretty amazing for me (certainly not the most popular kid in the school) when everybody was playing my game at lunch, after school, between classes - it was crazy. The school had to ban the game when keyboards started getting broken.
Another plus of a game assignment is that it's meaningful for kids of all skill levels. Many kids will write hangman, but someone's going to come in with a rendered first person shooter.
This article is old enough that probably nobody will ever read this, but anyway...
Swing does not use AWT for the "lowest level stuff". Most Swing classes derive from AWT components, but all the pixel pushing is done in Java, rather than with the OS widgets.
So Swing needed to specify the way all the widgets look. Aside from creating a great looking Look and Feel (Metal) Sun decided to make everything themable so that you could make Java apps look like Win32 apps or Mac apps or whatever. There is no doubt in my mind that all this generalness comes at a performance penalty. Take a look at the swing source sometime (in src.zip with the JDK) and note the incredible complexity. Undoubtedly stripping it out would result in a performance improvement. The question is whether the performance trade off is worth it.
Am I the only one who thinks that Sun over did it with all the pluggable look and feel stuff? It's definitely very cool to be able to click a button and have your application switch from Metal to Motif to Win32, but this comes at an enourmous complexity (and presumably speed) penalty. Try running a simple swing program through an analysis tool like OptimizeIt and you'll be horrified about how many classes need to get loaded for even the most simple app. I kind of wish they'd done a straight forward implementation of Metal (Which looks great in my opinion.)
This of course begs the question, how would NetPD even come up with a figure like this? If you want to get a list of everyone who's pirating Metallica material, it's pretty easy, you do a search for "Metallica" and maybe a few other variations and then pick all the user names out of the results. You could pretty easily devise an automated tool to do this, you run the search all weekend and presto, 300K odd names.
But claiming to know how many times a track was downloaded is a much more difficult problem. You have to somehow convince users to tell you how many people have downloaded a particular title. Obviously the server itself can do this because requests go through it, but as an outside client? Maybe the protocol lets you do this, but I doubt it.
Then, to find out how many unsigned artists were downloaded, you essentially have to track every client on the whole system and how many times every track was downloaded and then figure out which tracks were by unsigned artists. This essentially means having a master list of all signed artists in the world and doing some sort of text match against all the titles listed on Napster to eliminate signed artists. I find it highly unlikely that NetPD did this. I suppose you could develop a list of "known unsigned bands" (maybe scrape it from mp3.com or something) and see who downloads those tracks, but this is hardly accurate (And a great way to undercount.)
My guess is that this figure is invented. Whether NetPD or Lars invented it, I can't say.
It's the license (stupid)
on
Why Not MySQL?
·
· Score: 1
Much of the discussion so far has centered around the feature set, the fact that mySQL doesn't support the ACID properties of a database and whether or not that matters.
What hasn't yet been widely discussed is something that I would have expected the Slashdot community to care deeply about: the license. It's not open source. It's not even close. It's neither free as in beer, nor as in speech. Why does mySQL get a pass on this?
Have you used any of the Open Source databases like MySQL
There are multiple problems with the above statement. MySQL is neither free as in speech nor as in beer. Read their license for yourself. I have never understood why in a community where the nuances of a license are so critically important, mySQL seems to get a free pass.
The other problem is that mySQL does not have transaction support, and so calling it a database at all is a stretch at best. mySQL is a filesystem that uses SQL as a query language. Certainly this is useful, but comparing it to relational databases is dishonest. Of course mySQL is fast: It makes no attempt to ensure transaction semantics.
One could understand the community's attitude towards mySQL if no truly open source alternative existed, but PostgreSQL is both open source, and a "real" database.
So let me get this straight, Bleem destroys Sony's business model because they no longer have to sell you the hardware at a loss, but they can still sell you the high margin games, right?
It's almost like if I didn't buy an i-Opener but wanted to subscribe to the service using my own hardware.
This is quite simply not true. Modern Digital Watermarking is not visible to the human eye, but encoded in the low order bits of the color infomration of an image. They can be engineered to present a tradeoff between visibility, resistance to removal and number of bits for the embedded ID.
There are certainly some schemes that can withstand the image being cropped, resized, and even sometimes printed and rescanned. If you've used a copy of photoshop in the last couple of years, you should have seen it check every image you scan for an embedded watermark.
There are similar schemes for audio, and they don't necessarily require digital media. Again, there is a tradeoff between resistance to transformations, number of bits in the watermark and audibility.
Think about how much more bloated MS products will get when they have all the room on a DVD to play with. I just installed MS Visual Studio, and I thought it obscene that it required 550Mb. This is probably just the tip of the iceberg.
Slashdot Mirror
MIT runs a class called 6.033: Computer Systems Engineering. These lecture notes contain a list of projects that had great sums of money spent on them only to be abandoned. Also the reading list has a bunch of papers that discuss the "big splash" failures like Therac 25.
Go read back issues of Crypto-Gram and read up on all the lame hype laden "unbreakable" crypto schemes (often based on one time pads) that they destroy and then laugh at. If after reading all that you're still convinced you've got something, sure, go see a patent attorney.
My guess is you'll end up saving yourself the patent fees and a fair amount of humiliation by just letting it drop.
This is only slightly less optimistic than talking about installing a Mr. Fusion. The "on-board hydrogen storage system" (that fits in a six inch frame, no less!) is currently total fantasy. So there are a few things left to work out before we all get to drive cars like this.
Also needing to be addressed were the charges of systematic voter disenfranchisement, the polling places that closed early and so on.. but these are issues regardless of the implementation technology for the voting itself.
Let me describe the voting system Canada has: You register much as you do here. You show up at the polling place. They cross your name off the list and hand you a hard to forge ballot. You walk behind a little screen, put an X next to the person you want to vote for and stick it in a box. At the end of the day, representatives from each party and the media open the box and count the ballots. The results are delivered in a tree - local place reports to city, city probably to county, county to province. They add up all the results and they declare a winner.
Nothing about this fails to scale. In other words, a population 10x the size of Canada requires about 10x the number of volunteers which works out to be the same number of volunteers per capita.
This system seems so much more workable to me, there are so many fewer opportunities for breakdown.
- Is it Auditable? Yes, keep the ballots locked up and recount them.
- Is it anonymous? Yes, at least as much as touch screen voting.
- Is there any software / printers / touchscreens / whatever to fail? No.
Why do we need millions of dollars of development and plenty of technology to fail when a bunch of pieces of paper and some pens would do fine?I'd recommend you check out www.pcsforeveryone.com which is in the Cambridge MA area. They provide "near custom" services where you can put together pcs at almost the level of detail you would when doing it yourself (Like choosing which fan, which case, etc...) They're pretty clueful, know about Linux and can pre-install it, won't charge you for an OS if you don't want one, etc...
Even if you don't buy from them, it's worth reading their commentary on which parts are best and which parts work together / not with linux / etc...
Yes, it's using Unary, but you want to be able to compute "cost per byte" right? So somehow we need to be able to compare the value of this message to other binary encoded messages like the password for an Everquest account. This requires your message to encoded in binary.
However, if you want to be super picky, you can figure out the fractional requirement for how many bits are required to encode this message with this formula:
log (Number of states) / log (Number of states per symbol)
Which in this case works out to log 3 / log 2
or approximately 1.58 bits. Of course if this is the only message you're sending (which it was) you'd need to round up to the next integral number which brings us back to two.
From the ride of Paul Revere. So if you consider this three possible states (0-2 lanterns) it takes 2 bits to encode.
What was the "value" of that particular message? Up for debate of course but conventional wisdom is that the revolutionary war would have been lost without it and history would be very different.
One if by land, two if by sea.
The thing about Music is that nothing about it is _inherently_ expensive to produce. Sure, once you've thrown in the videos and the launch party and the services of the London Philharmonic it starts to get up there (Although I'd guess the services of the London Philharmonic are cheaper than you'd think) but the equipment, space and talent to just record music is generally within the range of every day people. Of course certain kinds of music are easier to record on the cheap than others (Moby can do it alone in his apartment because, well, he doesn't have any instruments) but with a nice Mac and $10K worth of extra hardware a talented bunch of people can put out some pretty respectable stuff. So music will live on, even if it's almost free. Same goes for books.
But what about Movies? Movies are going to be subject to the very same dynamic, although perhaps timeshifted a few years to the right. If the shit start to hit the movie industry, the world is going to start to look pretty different because movies are _expensive_. I mean even once you throw out the union pay scales and the staffing and the rules, blowing shit up (which is a staple of a lot of movies) is expensive; As are sets and crowds and animations and all the other stuff we see in our movies. Sure, you can still make "Clerks" and "The Blair Witch Project" pretty cheaply, but those aren't the only kinds of movies out there - not even an appreciable percentage if you're looking at Hollywood output. So the big budget movie could be a thing of the (soon to be) past.
Maybe movies will go all digital. Computing cycles will be so cheap and software so good that movies can be "filmed" at low cost by some Savant in his basement with the futuristic equivalent of an iMac and some Red Bull. But I wouldn't count on it.
It's a little more complicated than that. Christensen's contribution was that incumbent players miss disruptive technologies not because of bad management, but because of good management. The disruptive technologies initially are inferior, serve a smaller market and come with low margins and immature technology. At no point is it a rational decision to dump your successful high margin technology for an inferior product with a lower margin and a smaller market.
But what happens is that the disruptive technology improves with time to the point that it's "good enough" for your customers, even though it still can't do all the things the incumbent technology can do. And presto, the market disappears overnight.
I think a good example (although it's largely conjecture at this point) is the storage industry. EMC used to make a fortune selling extremely fast, reliable, high margin storage devices to enterprises. There have always been commodity PCs that did fileserving, but it was a different market: Less space, less reliable, less manageable, more failure prone! Who cares if it's cheaper.
Well over time, commodity PCs improve, technology gets better and people start to think, why not just have lots of commodity PCs and some clever software engineering stand in for my expensive EMC. Google launches and proves (to the people paying attention) that clusters of commodity PCs with clever software can be pretty damn fast and pretty reliable at storing rather a lot of data. And the cost is phenomenal.
At this point I'd say the writing is on the wall, but is EMC building commodity based storage solutions? Maybe in some tiny little skunkworks project, but not visibly. Even if they build it, will they launch it? Not clear. Why would they? They sell a Symmetrix for >$1M, the margins are huge on those things. And PCs are never going to be as fast as a tuned Symmetrix and fibre channel connection.
But you know what? Nobody cares.
So we'll see. Storage.
They just want a new reg key so they can get another 30 days free.
P2P research by well known research institutions is far from unheard of. MIT has Chord which is a project to produce robust scalable distributed systems using peer to peer ideas. They have an efficient hash lookup algorithm that could form the basis of many p2p systems and they have code available for download.
Of course they'll argue what they think the market effect will be, but then there's the simple question, "Didn't you guys say the same thing about the VCR?"
Actually they just launched a new version a couple of weeks ago - it's pretty cool and they seem to do a much better job of picking songs. It lists what all your friends are listening to, so if you have a couple of friends listening, someone is always listening to something cool and you can just jump right in with them.
1) An .xxx tld is set up so that porn companies can disitnguish themselves if they so choose.
2) AOL, libraries, schools and a bunch of others filter .xxx domains at the router.
3) Porn companies start getting pressure that they must register themselves in the .xxx domain. Eventually a law is passed.
4) Someone somewhere is tasked with deciding what qualifies under the mandatory .xxx law and what doesn't. While not being censored per se anything in the .xxx domain is totally marginalized because nobody can see it.
Is Mapplethorpe Porn? Is AIDS awareness? Is birth control information? What is porn? Let me guess, you'll know it when you see it right?
I think a real key here is that different students have vastly differing skill levels, so giving an assignment that can be handled meaningfully at all levels is great.
We did this when I was in highschool, and the most interesting (and gratifying) thing was the social impact. Video games were banned from our (and many other) High School computer labs, but games written by students were allowed.
I wrote a 4 player version of light cycles of tron, and it was pretty amazing for me (certainly not the most popular kid in the school) when everybody was playing my game at lunch, after school, between classes - it was crazy. The school had to ban the game when keyboards started getting broken.
Another plus of a game assignment is that it's meaningful for kids of all skill levels. Many kids will write hangman, but someone's going to come in with a rendered first person shooter.
Swing does not use AWT for the "lowest level stuff". Most Swing classes derive from AWT components, but all the pixel pushing is done in Java, rather than with the OS widgets.
So Swing needed to specify the way all the widgets look. Aside from creating a great looking Look and Feel (Metal) Sun decided to make everything themable so that you could make Java apps look like Win32 apps or Mac apps or whatever. There is no doubt in my mind that all this generalness comes at a performance penalty. Take a look at the swing source sometime (in src.zip with the JDK) and note the incredible complexity. Undoubtedly stripping it out would result in a performance improvement. The question is whether the performance trade off is worth it.
Am I the only one who thinks that Sun over did it with all the pluggable look and feel stuff? It's definitely very cool to be able to click a button and have your application switch from Metal to Motif to Win32, but this comes at an enourmous complexity (and presumably speed) penalty. Try running a simple swing program through an analysis tool like OptimizeIt and you'll be horrified about how many classes need to get loaded for even the most simple app. I kind of wish they'd done a straight forward implementation of Metal (Which looks great in my opinion.)
This of course begs the question, how would NetPD even come up with a figure like this? If you want to get a list of everyone who's pirating Metallica material, it's pretty easy, you do a search for "Metallica" and maybe a few other variations and then pick all the user names out of the results. You could pretty easily devise an automated tool to do this, you run the search all weekend and presto, 300K odd names.
But claiming to know how many times a track was downloaded is a much more difficult problem. You have to somehow convince users to tell you how many people have downloaded a particular title. Obviously the server itself can do this because requests go through it, but as an outside client? Maybe the protocol lets you do this, but I doubt it.
Then, to find out how many unsigned artists were downloaded, you essentially have to track every client on the whole system and how many times every track was downloaded and then figure out which tracks were by unsigned artists. This essentially means having a master list of all signed artists in the world and doing some sort of text match against all the titles listed on Napster to eliminate signed artists. I find it highly unlikely that NetPD did this. I suppose you could develop a list of "known unsigned bands" (maybe scrape it from mp3.com or something) and see who downloads those tracks, but this is hardly accurate (And a great way to undercount.)
My guess is that this figure is invented. Whether NetPD or Lars invented it, I can't say.
What hasn't yet been widely discussed is something that I would have expected the Slashdot community to care deeply about: the license. It's not open source. It's not even close. It's neither free as in beer, nor as in speech. Why does mySQL get a pass on this?
There are multiple problems with the above statement. MySQL is neither free as in speech nor as in beer. Read their license for yourself. I have never understood why in a community where the nuances of a license are so critically important, mySQL seems to get a free pass.
The other problem is that mySQL does not have transaction support, and so calling it a database at all is a stretch at best. mySQL is a filesystem that uses SQL as a query language. Certainly this is useful, but comparing it to relational databases is dishonest. Of course mySQL is fast: It makes no attempt to ensure transaction semantics.
One could understand the community's attitude towards mySQL if no truly open source alternative existed, but PostgreSQL is both open source, and a "real" database.
It's almost like if I didn't buy an i-Opener but wanted to subscribe to the service using my own hardware.
Hardly a business model buster.
This is quite simply not true. Modern Digital Watermarking is not visible to the human eye, but encoded in the low order bits of the color infomration of an image. They can be engineered to present a tradeoff between visibility, resistance to removal and number of bits for the embedded ID.
There are certainly some schemes that can withstand the image being cropped, resized, and even sometimes printed and rescanned. If you've used a copy of photoshop in the last couple of years, you should have seen it check every image you scan for an embedded watermark.
There are similar schemes for audio, and they don't necessarily require digital media. Again, there is a tradeoff between resistance to transformations, number of bits in the watermark and audibility.
Think about how much more bloated MS products will get when they have all the room on a DVD to play with. I just installed MS Visual Studio, and I thought it obscene that it required 550Mb. This is probably just the tip of the iceberg.