AutoStitch is a research project from the U of BC, but the demo app is advanced enough that you can just drop a dozen images into the thing and it figures out the rest. Much easier than any other tool I've used.
I will say that the quality of a fully-HD show is enough to make me watch it live, even the commercials. Some of the rebroadcast movies (Sound of Music, Pixar flicks, Rudolph) are really worth seeing. But it's just sad when you realize how much potential is lost in the average HD broadcast. Very few content is shot in 16:9, even fewer in anything higher than 480p (PBS is the exception, but seems to go out when a truck passes by my house.)
What bugs me the most is when a show/commercial/movie preview is a letterboxed 4:3 format -- you end up with a small rectangle inside of the $2000 rectangle that is your HD-capable TV. I also don't realize why so few advertisers take advantage of the format. Do the broadcasters charge by the MB? I'll watch *any* well-shot HD commercial.
This is why I think HD-DVD will fail -- if consumers don't even demand hi-fi TV, why would they demand hi-fi DVDs?
You have two kind of people in a company: People whose job it is to be skeptical (engineers) and people whose job it is to be optimistic (sales/marketing). The former think in terms of logic, the second think in terms of perception. That's why sales guys writing code doesn't make sense.
The idea of sales guys looking at code scares the crap out of me.
I loved eMusic when it was in the all-you-can-eat for $9.95 or something ridiculous per month.. most of my mp3 collection is from that year-long downloading orgy. I've been exposed to more eclectic music than I'd ever get looking for torr3ntz and such.
Unfortunately, in that time I've downloaded all the Pavement I'd ever want:) Maybe I'll give them a look, if the cap is really 90 songs per month.
If you want to limit your career options to database-driven web apps or system scripting, sure -- go ahead and forget the math. Just remember that computers are fancy calculators, after all, and they are all about the math.
OTOH, if you want to be known as the guy who solves the tough problems, you should learn the math. For instance, my current job involves dealing with all kinds of power-electronics measurements and figuring out things like the dynamics of battery cells. Since I have some mathematical background, I can interpret papers and derive equations to solve some of my problems.
The things I don't quite know, like Kalman filters, I can kind of understand and decide whether or not we need an expert on the subject, because I've come across similar concepts in the past. I know about different types of filtering from audio processing theory, and I can draw those parallels to understand other systems. I can even talk about "integrals" and impress people, when I know it's just the area under a curve.
I don't think it's a question of "computer scientist" vs "software engineer" either. An engineer leverages their knowledge of math and science to solve real problems. A scientist leverages existing knowledge to derive more knowledge.
It's a question of being a bad-ass engineer vs. a code monkey:)
Beware first generation hardware, and if you decide to load the chamber and pull the trigger, get Applecare. I went through *three* iBook G3's before Apple replaced it with a shiny new iBook G4. Nice of them, but I would really rather have my 40 hours of lost productivity back.
Also, consider that the Lunar Module and Space Shuttle both have autopilots, yet neither has been allowed to operate fully until touchdown. The main reasoning behind this is that pilots want to be "in the loop" during those last critical seconds before touchdown, and if something fails the pilot wants to be fully engaged at the time of failure, and not have to switch between monitoring-the-autopilot mode to flying-the-vehicle mode.
Couple observations from using C++ in embedded soft-real-time systems (and games):
* The smaller your subset of C++, the more portable. You can make a perfectly functional C++ app without using operator overloading, RTTI, exceptions, even virtual functions.
* The smaller your subset of C++, the less bugs that you'll blame on the compiler but will turn out to be your misinterpretation of the language (this applies unless you're Stroustrup, or maybe the author of more than one C++ compiler)
* Stack allocations and constructors/destructors are good.
* Little features like 'const' and references (basically pointers that you can't re-assign and can't be NULL) can be handy.
* STL is handy in small doses, but don't overdo it. You'll get error messages that no mortal can parse.
I work with embedded devices. I loves me some wool pants, but just try poking around electronics after strolling around a dry room. Bzzzzap.
Pie in the Sky Question
on
Ask Sid Meier
·
· Score: 1
What is the game you've always wanted to design, but feel you won't in your lifetime because of technological constraints, market constraints, or just plain old haven't-figured-out-the-right-approach constraints?
This is dumb. DST has been in effect for about 40 years. It precedes the software industry as we know it. The cost of patching all the devices, hardware + software, is likely to run into the billions. Maybe these congressmen think our IT industry needs a quick jolt? Great, but what about the expense to the industries using the technology.
How about initiatives for smart energy policies instead of quick hacks? Almost every household in the U.S. could save 1% on their energy bill if they just had a little education/incentive.
I say we have a national referendum on DST, and settle this once and for all. My bet is it would be sent packing, by a landslide.
Yes, innovation will suffer because the smaller shareware/freeware/demoware players will be fed up with the whole patent business and go raise goats or something... while the big players trade patents with one another to make the same three games they've made for the past ten years. The big boys won't even have the little guys to copy ideas from! EA, Sony, and Microsoft are no doubt salivating at the thought of complete idea-monopoly of whatever the next big genre is (MMOFPSRPGRTSSimChess?).
You are in a maze of twisty little legal passages, all alike.
I've got a website that does something similar, but for two arbitrary Wikipedia entries. For instance, consider the case of Bin Laden to Henry Kissinger. The paths go through Cyrus Vance, Christopher Hitchens, and Donald Rumsfeld. Does this mean anything? Probably not. Are there longer paths that might reveal some sort of conspiracy? Probably not, but there are probably other, longer paths that might reveal more interesting relationships. I'd like to see what kind of metric they use for "interestingness", especially given the amount of data they plan to incorporate.
What would happen if you ran such a thing against a Lexis-Nexus-sized database?
Why don't you *both* take the money you're spending talking about what else to do with the money we're spending on the shuttle, and use it to set up aerospace companies in third-world countries to build your *own* space shuttle, and then fly it into space, and then come back and post your adventures. Yeah.
Ya know, I don't begrudge the screenwriters (at least not until having seen the movie). Movies based on books (or radio plays) shouldn't be expected to lift their dialogue line-for-line. They're different mediums and have different expectations. HHGG is at its heart a story, and can have many manifestations, just as Shakespeare's plays have successfully been transplanted to other times and places.
We have all kind of "AI-like" technology in our computers right now -- spam filtering, intelligent search engines, collaborative filtering (for instance TiVo recommendations), speech/image/OCR/handwriting recognition, etc. This stuff is real and useful and improving all the time. We just don't call it "AI" as much, because "AI" is a word associated with failed aspirations. What we have are highly refined statistical systems that are optimized for a particular problem.
What the "baby bootstrap" is really referring to is "the great emergent AI" which, like HAL-9000, will be able to empathize with humans, navigate a starship, and play a mean game of chess -- because if a system can perform one intelligent operation, it can perform another operation requiring an equal amount of intelligence, right?
One major stumbling block (I think) is that of optimization. The relatively simple problem of speech recognition takes a major percentage of a modern CPU's power, and is still 95-98% accurate. This is heavily optimized software written by very smart people with a couple decades of research behind it.
A hypothetical "great emergent AI" system would have to perform the function of speech-recognition -- since it is supposed to be like a child or like a HAL-9000 -- but it would have to come up with a same-or-better implementation of this very complex algorithm, using some emergent process. It would have to figure out the equivilent of FFTs, cepstral coefficients, lattice search... stuff that isn't instantly derivable from a + b = c.
What we think our brain does is solve problems with a semi-brute-force algorithm. (Just throw billions of neurons at it!) However we still don't have the kind of computing power to implement a one-algorithm-fits-all learning process like the brain. Unfortunately, research for this "generic learning" is in a rut, with genetic algorithms and neural networks being exhausted top contenders. What will be next?
I am not a J2EE fanboy, but I depend on it for my bread and butter. As long as I work within the federal/enterprise development sphere, I will probably depend on it for quite some time. This is because big organizations in the U.S. are standardizing on J2EE, and to some extent dot-NET.
One reason for this is because their IT departments discovered a few years ago that they have accumulated all sorts of crufty code written against multiple languages and platforms, and had no way to manage it because all the contractors who wrote it had taken the money and run. Seeing shrinking IT budgets, and new platforms (like ROR) coming out every year, they decided to make their lives easier by standardizing on the most best-of-breed (read: most resumes available) platforms. This would allow their IT folks to only have to learn one set of technologies to deploy and manage applications internally.
IT managers like standard platforms, because they are sold on the idea that they can develop "web services" which can let apps "transparently share data" and allow for "collaborative whatchamacatchie". Most often, the reality is that these organizations still have a tremendous amount of data scattered across N databases and M organizations, and no magic dev platform is going to solve that without effort.
I regard most of the J2EE architecture as over-engineered and overkill on most projects I've done. I prefer to use it stripped-down as I can get away with. Heck, I could probably embed a Jython interpreter in a servlet and use it on a project without anyone noticing. But for large organizations, the trend is toward less diversity.
It's now Non-Technical Professional Appreciation Day. Let me list the things that I'm glad my PHB/business-dev/sales folks get to do:
* Go to project meetings and listen to client's problems for hours on end * Fly halfway around the world to drum up new business * Resolve sticky contractual issues that involve many days of phone calls, faxes, and lawyers * Translate for/run defense against upper management
There is no question that some organizations are top-heavy and it slows them down. This is why we have the free enterprise system, to allow the nimble to outsmart the slow. But I don't see it getting worse. In fact we now have the technology and management theory to make big enterprise more efficient than ever -- witness the Wal Mart phenomenon.
If you complain about anything, complain about the government's policies not making the playing field level enough. But I don't agree with your worker's revolution thesis. As long as there are people willing to give up freedom for security, there will be the employer-employee agreement.
Oh BTW if this is a troll, please ignore -- I gave up my mod points anyway:)
How does a bunch of spanish speaking illegal immigrants write a better technical report than MIT students?
One can assume that for the Spanish speaking folks, English is a second language. For the MIT student, English is fifth or sixth, right behind FORTH...
Weren't there services that did this in the '90's, like Blink.com? They also did auto-recommendation, kinda nifty. Of course back in the day we didn't have all the nifty Moz plugins, they had to use Java craplets, so not as useful.
I've had this idea forever, but what I really wanna see out of these things (besides transparent, reliable bookmark syncing) is auto-categorization. I'm too lazy to put things into appropriate folders.
I dimly remember that ad. What comes to mind first is Borland's ObjectVision but that may be too early for '96. OV was kind of an early competitor to VB.
Slashdot Mirror
AutoStitch is a research project from the U of BC, but the demo app is advanced enough that you can just drop a dozen images into the thing and it figures out the rest. Much easier than any other tool I've used.
I will say that the quality of a fully-HD show is enough to make me watch it live, even the commercials. Some of the rebroadcast movies (Sound of Music, Pixar flicks, Rudolph) are really worth seeing. But it's just sad when you realize how much potential is lost in the average HD broadcast. Very few content is shot in 16:9, even fewer in anything higher than 480p (PBS is the exception, but seems to go out when a truck passes by my house.)
What bugs me the most is when a show/commercial/movie preview is a letterboxed 4:3 format -- you end up with a small rectangle inside of the $2000 rectangle that is your HD-capable TV. I also don't realize why so few advertisers take advantage of the format. Do the broadcasters charge by the MB? I'll watch *any* well-shot HD commercial.
This is why I think HD-DVD will fail -- if consumers don't even demand hi-fi TV, why would they demand hi-fi DVDs?
You have two kind of people in a company: People whose job it is to be skeptical (engineers) and people whose job it is to be optimistic (sales/marketing). The former think in terms of logic, the second think in terms of perception. That's why sales guys writing code doesn't make sense.
The idea of sales guys looking at code scares the crap out of me.
I loved eMusic when it was in the all-you-can-eat for $9.95 or something ridiculous per month .. most of my mp3 collection is from that year-long downloading orgy. I've been exposed to more eclectic music than I'd ever get looking for torr3ntz and such.
:) Maybe I'll give them a look, if the cap is really 90 songs per month.
Unfortunately, in that time I've downloaded all the Pavement I'd ever want
If you want to limit your career options to database-driven web apps or system scripting, sure -- go ahead and forget the math. Just remember that computers are fancy calculators, after all, and they are all about the math.
:)
OTOH, if you want to be known as the guy who solves the tough problems, you should learn the math. For instance, my current job involves dealing with all kinds of power-electronics measurements and figuring out things like the dynamics of battery cells. Since I have some mathematical background, I can interpret papers and derive equations to solve some of my problems.
The things I don't quite know, like Kalman filters, I can kind of understand and decide whether or not we need an expert on the subject, because I've come across similar concepts in the past. I know about different types of filtering from audio processing theory, and I can draw those parallels to understand other systems. I can even talk about "integrals" and impress people, when I know it's just the area under a curve.
I don't think it's a question of "computer scientist" vs "software engineer" either. An engineer leverages their knowledge of math and science to solve real problems. A scientist leverages existing knowledge to derive more knowledge.
It's a question of being a bad-ass engineer vs. a code monkey
Beware first generation hardware, and if you decide to load the chamber and pull the trigger, get Applecare. I went through *three* iBook G3's before Apple replaced it with a shiny new iBook G4. Nice of them, but I would really rather have my 40 hours of lost productivity back.
Also, consider that the Lunar Module and Space Shuttle both have autopilots, yet neither has been allowed to operate fully until touchdown. The main reasoning behind this is that pilots want to be "in the loop" during those last critical seconds before touchdown, and if something fails the pilot wants to be fully engaged at the time of failure, and not have to switch between monitoring-the-autopilot mode to flying-the-vehicle mode.
Couple observations from using C++ in embedded soft-real-time systems (and games):
* The smaller your subset of C++, the more portable. You can make a perfectly functional C++ app without using operator overloading, RTTI, exceptions, even virtual functions.
* The smaller your subset of C++, the less bugs that you'll blame on the compiler but will turn out to be your misinterpretation of the language (this applies unless you're Stroustrup, or maybe the author of more than one C++ compiler)
* Stack allocations and constructors/destructors are good.
* Little features like 'const' and references (basically pointers that you can't re-assign and can't be NULL) can be handy.
* STL is handy in small doses, but don't overdo it. You'll get error messages that no mortal can parse.
I work with embedded devices. I loves me some wool pants, but just try poking around electronics after strolling around a dry room. Bzzzzap.
What is the game you've always wanted to design, but feel you won't in your lifetime because of technological constraints, market constraints, or just plain old haven't-figured-out-the-right-approach constraints?
Why don't you just add one more function to your SOAP server and have your exception handler connect to that?
You see, these SOAP servers go to eleven.
This is dumb. DST has been in effect for about 40 years. It precedes the software industry as we know it. The cost of patching all the devices, hardware + software, is likely to run into the billions. Maybe these congressmen think our IT industry needs a quick jolt? Great, but what about the expense to the industries using the technology.
How about initiatives for smart energy policies instead of quick hacks? Almost every household in the U.S. could save 1% on their energy bill if they just had a little education/incentive.
I say we have a national referendum on DST, and settle this once and for all. My bet is it would be sent packing, by a landslide.
Yes, innovation will suffer because the smaller shareware/freeware/demoware players will be fed up with the whole patent business and go raise goats or something ... while the big players trade patents with one another to make the same three games they've made for the past ten years. The big boys won't even have the little guys to copy ideas from! EA, Sony, and Microsoft are no doubt salivating at the thought of complete idea-monopoly of whatever the next big genre is (MMOFPSRPGRTSSimChess?).
You are in a maze of twisty little legal passages, all alike.
I've got a website that does something similar, but for two arbitrary Wikipedia entries. For instance, consider the case of Bin Laden to Henry Kissinger. The paths go through Cyrus Vance, Christopher Hitchens, and Donald Rumsfeld. Does this mean anything? Probably not. Are there longer paths that might reveal some sort of conspiracy? Probably not, but there are probably other, longer paths that might reveal more interesting relationships. I'd like to see what kind of metric they use for "interestingness", especially given the amount of data they plan to incorporate. What would happen if you ran such a thing against a Lexis-Nexus-sized database?
Why don't you *both* take the money you're spending talking about what else to do with the money we're spending on the shuttle, and use it to set up aerospace companies in third-world countries to build your *own* space shuttle, and then fly it into space, and then come back and post your adventures. Yeah.
Yeah, he's probably right. Caves get *terrible* receiption. You'd have to step outside and face the leopards just to send a single IM!
(P.S. I know not many people lived in caves 200 years ago, etc...)
Ya know, I don't begrudge the screenwriters (at least not until having seen the movie). Movies based on books (or radio plays) shouldn't be expected to lift their dialogue line-for-line. They're different mediums and have different expectations. HHGG is at its heart a story, and can have many manifestations, just as Shakespeare's plays have successfully been transplanted to other times and places.
We have all kind of "AI-like" technology in our computers right now -- spam filtering, intelligent search engines, collaborative filtering (for instance TiVo recommendations), speech/image/OCR/handwriting recognition, etc. This stuff is real and useful and improving all the time. We just don't call it "AI" as much, because "AI" is a word associated with failed aspirations. What we have are highly refined statistical systems that are optimized for a particular problem.
... stuff that isn't instantly derivable from a + b = c.
What the "baby bootstrap" is really referring to is "the great emergent AI" which, like HAL-9000, will be able to empathize with humans, navigate a starship, and play a mean game of chess -- because if a system can perform one intelligent operation, it can perform another operation requiring an equal amount of intelligence, right?
One major stumbling block (I think) is that of optimization. The relatively simple problem of speech recognition takes a major percentage of a modern CPU's power, and is still 95-98% accurate. This is heavily optimized software written by very smart people with a couple decades of research behind it.
A hypothetical "great emergent AI" system would have to perform the function of speech-recognition -- since it is supposed to be like a child or like a HAL-9000 -- but it would have to come up with a same-or-better implementation of this very complex algorithm, using some emergent process. It would have to figure out the equivilent of FFTs, cepstral coefficients, lattice search
What we think our brain does is solve problems with a semi-brute-force algorithm. (Just throw billions of neurons at it!) However we still don't have the kind of computing power to implement a one-algorithm-fits-all learning process like the brain. Unfortunately, research for this "generic learning" is in a rut, with genetic algorithms and neural networks being exhausted top contenders. What will be next?
I don't know -- any AC post with the subject "You are all wrong" does not merit my mod-up.
I am not a J2EE fanboy, but I depend on it for my bread and butter. As long as I work within the federal/enterprise development sphere, I will probably depend on it for quite some time. This is because big organizations in the U.S. are standardizing on J2EE, and to some extent dot-NET.
One reason for this is because their IT departments discovered a few years ago that they have accumulated all sorts of crufty code written against multiple languages and platforms, and had no way to manage it because all the contractors who wrote it had taken the money and run. Seeing shrinking IT budgets, and new platforms (like ROR) coming out every year, they decided to make their lives easier by standardizing on the most best-of-breed (read: most resumes available) platforms. This would allow their IT folks to only have to learn one set of technologies to deploy and manage applications internally.
IT managers like standard platforms, because they are sold on the idea that they can develop "web services" which can let apps "transparently share data" and allow for "collaborative whatchamacatchie". Most often, the reality is that these organizations still have a tremendous amount of data scattered across N databases and M organizations, and no magic dev platform is going to solve that without effort.
I regard most of the J2EE architecture as over-engineered and overkill on most projects I've done. I prefer to use it stripped-down as I can get away with. Heck, I could probably embed a Jython interpreter in a servlet and use it on a project without anyone noticing. But for large organizations, the trend is toward less diversity.
Huh?
:)
It's now Non-Technical Professional Appreciation Day. Let me list the things that I'm glad my PHB/business-dev/sales folks get to do:
* Go to project meetings and listen to client's problems for hours on end
* Fly halfway around the world to drum up new business
* Resolve sticky contractual issues that involve many days of phone calls, faxes, and lawyers
* Translate for/run defense against upper management
There is no question that some organizations are top-heavy and it slows them down. This is why we have the free enterprise system, to allow the nimble to outsmart the slow. But I don't see it getting worse. In fact we now have the technology and management theory to make big enterprise more efficient than ever -- witness the Wal Mart phenomenon.
If you complain about anything, complain about the government's policies not making the playing field level enough. But I don't agree with your worker's revolution thesis. As long as there are people willing to give up freedom for security, there will be the employer-employee agreement.
Oh BTW if this is a troll, please ignore -- I gave up my mod points anyway
How does a bunch of spanish speaking illegal immigrants write a better technical report than MIT students?
One can assume that for the Spanish speaking folks, English is a second language. For the MIT student, English is fifth or sixth, right behind FORTH...
Weren't there services that did this in the '90's, like Blink.com? They also did auto-recommendation, kinda nifty. Of course back in the day we didn't have all the nifty Moz plugins, they had to use Java craplets, so not as useful.
I've had this idea forever, but what I really wanna see out of these things (besides transparent, reliable bookmark syncing) is auto-categorization. I'm too lazy to put things into appropriate folders.
I dimly remember that ad. What comes to mind first is Borland's ObjectVision but that may be too early for '96. OV was kind of an early competitor to VB.
Hard *bought*, that is:
c qu isitions
http://en.wikipedia.org/wiki/List_of_Symantec_a
Gobble, gobble. And the patents are part of each deal. Symantec (like most tech companies) also has an aggressive internal patent search process.