Bottom line is that you need to know legal precedent (i.e how courts have applied copyright law in cases similar to one you may be interested in). Copyright only applies to "creative" works, where creativity is interpreted quite broadly.
In the case of a photo of a statue, copyright law does apply (the photo is considered a creative work), and copyright of the photo therefore belongs to the photographer (the photo being a seperate creative work than the scultpure itself). If multiple photographers took photos of it, each would own copyright to their own work.
However, in the case of a photo of a painting (vs statue), courts have previously found that copyright law does not apply since they've ruled that such a photo is a "slavish copy" lacking in any creativity. I doubt we've heard the end of such lawsuits though. One photo might be akin to a photocopy, but another might employ lighting, angle, composition, etc, and meet the (rather weak) standard of creativity that copyright law requires.
Software writing is a highly creative process, and most if not all software is going to be covered by copyright. If you just provided a software contractor with functional requirements specifications and left him to design and write the code, then you've not in any way removed the need for creativity (to do that you've had to have designed and implemented the code yourself!), so copyright law would indeed apply and as always the author of the creative work (programmer) by default owns the copyright.
If you wanted to contract for a piece of software to be written and you wanted to be granted certain distribution rights to it, then you'd need to specify exactly what rights it is that you are purchasing in the purchase contract. By default (if you don't put anything explicit in writing) then the programmer retains all rights. You may not like it, but that's the law.
If you feel that your requirements specification contains some valuable intellectual property that you'd like to protect, then you'd have to look into appropriate mechanisms to do that (trade secret, patent if applicable, etc), but that's entirely unrelated to who owns the copyright to a creative work based on those specifications.
If I paid someone to paint a painting for me, I own the painting.
Sort of... you own the physical painting, but that isn't the same as owning the copyright. Whether you owned the copyright (in whole or specific rights of duplication, etc) would depend on the purchase contract you arranged with the painter.
Same with a book. If you buy a book you own the physical object, but you don't own the copyright. You can't go making copies of the book and selling them yourself.
Same with software. If you "buy" a piece of software, what you've bought depends on the contract. You're typically going to be buying the physical media, and rights to use it either indefinitely or maybe for a limited period of time (cf renting a video), but not any copyright ownership unless that was specified in the purchase contract.
It doesn't make any difference whether you're buying an off-the shelf painting or software title, or comissioning someone to create a custom work for you. What you end up owning and the rights you have depend on the contract. Copyright law says that the author by default owns the copyright, so if you want anything different you need to put it in the contract.
There's no difference between paintings, software, books or any other type of copyrightable material - the same copyright laws cover them all.
Speech recognition technology is really still in its infancy... it's possible to get good results but only under the most controlled of circumstances... high quality microphone, no background noise, clear diction, recognition engine trained for the speaker, etc. Even then it may depend on what you're actually saying, since in the case of any ambiguity a smart recognition engine will fall back to grammatical analysis and word frequency counts etc to try to guess right.
The real problem is that speech recognition requires artificial intelligence to do right and we don't have it. Often we understand speech that word for word is basically unintelligible, but we automatically apply context and intelligence to figure out what the speaker was trying to enunciate. Without full AI, a computer can't do that - it's much more limited.
When the foreigner serving you at McDonalds mumbles something after taking your order, you only need to understand a single word, or maybe not even that, to realize they are saying "to stay or take out?", but a computer today would need the person to have spoken clearly enough to have made out the words. Ditto for a stream-of-consciousness rambling GMail voice message with highway noise in the background - YOU may be able to FIGURE OUT what is being said, but that's not the same as the words actually being intelligible which is what a computer, without AI, would need to be able to transcribe it accurately.
Having a clear speech signal in the first place, or having a broadband vs telephone limited one, isn't going to make much difference, except under otherwise very controlled conditions.
No - look at the data from top of TFA (those two tables). IE was in 5th place 50% of the time vs what should have been 20% if it had been random..
It was a large statistically meaningful sample. The problem here isn't round-off or a poor random number generator. It's using an invalid list randomize algorithm.
Getting random numbers evenly distributed over an arbitrary range isn't a problem as long as you've got them available evenly distributed over *some* range. You just scale them.
No - look at the example of the two-element list in the link I provided. I'll copy it here for convenience:
A commonly used shuffle algorithm attaches random tags to elements of a sequence and then sorts the sequence by the tags. Although this algorithm performs well in practice, it does not achieve the perfect random shuffling.
Let us consider the simplest example (which happens to be the worst case): a sequence of two elements, [a b]. According to the shuffle-by-random-keys algorithm, we pick two binary random numbers, and associate the first number with the 'a' element, and the second number with the 'b' element. The two numbers are the tags (keys) for the elements of the sequence. We then sort the keyed sequence in the ascending order of the keys. We assume a stable sort algorithm. There are only 4 possible combinations of two binary random numbers. Therefore, there are only four possible tagged sequences:
[(0,a) (0,b)]
[(0,a) (1,b)]
[(1,a) (0,b)]
[(1,a) (1,b)] After the sorting, the sequences become:
[(0,a) (0,b)]
[(0,a) (1,b)]
[(0,b) (1,a)]
[(1,a) (1,b)]
As we can see, after the sorting, the sequence [a, b] is three times more likely than the sequence [b, a]. That can't be a perfect shuffle.
**
The example is assuming that of the two random numbers, a b - not a == b.
Never mind the theory, look at the results from TFA. IE was in 5th place (out of 5) 50% of the time in a large statistically meaningful result set.
Well, as it happens (if you read TFA) Microsoft's solution ends up being biased against Microsoft rather than for them. It hard to say which is worse, though; in one case you work against your own market share, in the other you end up wasting time and money as the EU hauls you back in court for unfair practices.
It seems the least cost option for Microsoft would have been to have gotten this right first time (as is always the case with software).
It's not an issue of how to get a truly random number, or of seeding a random number generator, but rather how to you use a source of random numbers to randomly order a list.
Some "reasonable sounding" methods don't actually work - e.g. attach random numbers to each list item and sort the list by these numbers (1). Microsoft used a similar method of sorting using a random comparator.
Some simple methods that DO work are picking a random permutation or executing a bunch of random swap operations on the list.
Hard to say whether that's better or worse than the common NASA "cost plus" contract where they agree to pay development costs plus an agreed profit margin. Where's the incentive for the developer to keep cost under control?
I worked at Acorn in the early 80's and knew Steve Furber, but it's amusing to think that most of the Slashdot crowd probably wasn't even born then. I assume the average age here nowadays must be college age or thereabouts.
Not surprising they don't know what ARM originally stood for.
If NeXT sold to Apple for $400M, why did NeXT die?
It doesn't make much difference from the point of view of success, which is what we're discussing. Bottom line is he started a company and sold it it for $400M.
Presumably you know as well as me why NeXT (the company vs the software) died... I'd put it down to the difficulty of trying to get momentum for a brand new operating system.
I find that hard to believe, given the small-minded nature of your potshots, but if it's true then bully for you. You do realize that NeXT software is still in use too, right? (Apple Cocoa, Objective-C).
Maybe "one hit wonder" is inaccurate, but "cult leader" may be sufficient.
He has what it takes to motivate people to success. No need to be disparaging about it in this context.
I wouldn't put Apple in that category... Look at Steve Jobs' track record and it's pretty amazing...
Apple, NeXT (sold to Apple for $400M+), Pixar, Apple again (Jobs return and resuscitation of the company via iMac, iPod, iPhone...).
That's a quite of a string of major successes, even if there have been a few product failures at Apple along the way. I'm not sure that Jobs himself has had many start-up failures.
So, it's not right to call Apple/Jobs a one-hit wonder. Maybe he's just serially lucky (SOMEONE has to be at the extremes of the bell curve of any phenomemon), maybe not, but that's a different story.
It's not that failure is good for you (although of course you can learn from it), but rather that it's pretty much inevitable, so you better learn how to plan for and deal with it.
For example, the success rate for start-up companies is quite small (10% - I forget), so if you're going to try a start-up it's best not to commit yourself to such a degree that it hurts your ability to shake off the failure and try again.. and again..
There's an interesting book about the start-up experience of AutoDesk (the company that created AutoCAD) called "The AutoDesk File" by John Walker, that says the same thing. AutoDesk's founders never expected to start a CAD software company... but in the end that was the product idea that became successful. The general conclusion was keep trying and let marketplace success not preconceived ideas dictate your level of financial/etc commitment.
There's plenty of reasons to be suspect of WikiPedia, not least that officially it doesn't even strive for the truth - just for verifiyability (basically a published source).
However, there have been studies done showing that WikiPedia articles are on average just as accurate as Encyclopedia Britannica ones - both have similar average numbers of errors per article.
A large percentage of older programmers are unable to learn a new programming model. For example: Object Oriented coding.
I hate to break it to ya, sunshine, but OO design isn't anything new, even if languages to make it easy are more so.
Want OO design and polymorhism in 'C' or assembler? Just use a struct with pointers to functions. Things like implicitly called constructors and destructors are just syntactic sugar.
Us old fucks aren't quite as stupid as you imagine us to be!;-)
For simple projects that's true - where the application itself is straightforward and you just need people conversant in the tools you need to use to build it. That's not why you hire highly experienced people. No technology really takes more that 5 years to thoroughly master (although some complex domains may do).
Where you DO need experienced people is where the application is significantly complex or critical, where what the experience buys you is expertise in software architecture, and design. There are many obvious high-profile jobs like this - compiler design, radar systems, speech recognition, etc, but also many much more mundane ones... complex multi-threaded realtime software at the core of communications systems, industrial control systems, etc.
For these types of projects what you really need is years of experience in designing big complex systems. Being a hotshot in C++ or Java or the technology du jour (however many/few years it took you to get there) doesn't help at all if the scope and complexity of the project is just way over your head. This doesn't just apply to massive software projects - there are plently of complex one-man projects where someone without significant design experience is going to be struggling, making mistakes, and making slow progress (and ending up with a bunch of unmaintainable code that reflects the struggles) as opposed to a more senior person who'll design it right and in a fraction of the time because the job was well below his experience level rather than above it.
For hobby use, with no particular application in mind, I'd suggest learning Python. If you enjoy that and want to get more serious about programming, then learn Java, C# or C++. If you have in mind WHAT you'd like to program, then different advice may apply.
Python is well designed and scales up reasonably well (unlike, say, Perl, which does neither). It's more fun to start with an interpreted language due to the instant feedback you get.
With COBOL you may have a point - OTOH are there really any full-time COBOL programmers out there? I thought they all went back into retirement after coming out for the Y2K feeding frenzy!
OTOH, say I'm looking to hire someone to do Java development... I'd MUCH prefer a C++ programmer with 10 years experience over a fresh graduate who knows Java. The C++ programmer will learn Java few months, but the newbie will need 10 years to get 10 years worth of programming experience.
The difference in productivity between a programmer with a lot of experience and a newbie can be astounding... A seasoned professional who puts his years of experience to play in designing it right can do in a week what a newbie may struggle months to get working (and still end up with an unmaintainable poorly written mountain of code vs the clean solution of the pro).
It seems that functionally Clutter has been superceded by advances in Qt.. Qt supports OpenGL/ES/VG backends and has a new "declarative UI" for designing animated and custom UIs. In fact since the Nokia aquisition Qt seems to have all but forgotten about the desktop and most new features are squarely aimed at implementing fluid custom iPhone-like interfaces on smartphone/netbook targets.
Looking at that Fisherman's Wharf in Google maps some more, in general they are just presenting a menu of geo-tagged photos, but the big round Wharf sign itself is a special case.... For that they've grouped a bunch of photos of the same object and have determined how these map onto a street view photo of the same object.
The fact that no other user photos (in this scene at least) are mapped to street view makes it appear this was done as a special case. Maybe the user photos were hand tagged as being of the same object? In any case, this is just mapping to a master photo, not integrating into a 3-D scene.
The user photos are not overlaid on the street view (neither were they in the Microsoft demo) - they are just presented as an alternative view of the same location. i.e. Google is just using geo-tags to give to a user-selectable bunch of photos taken at/near that location. Yawn.
That is completely unrelated to what Microsoft's demo did. It didn't just give you a menu of photos geo-tagged to the location, or simply overlay them at that spot. It actually recognized the content of the photos, where they belong in the 3-D scene, and incorporated them into the 3-D scene in augmented reality fashion. Way cool!
Slashdot Mirror
We're talking about copyright, not subject permission. Nice try to change the subject.
Yes, legal precedent has do be from the same legal jusisdiction. Well, duh!
Bottom line is that you need to know legal precedent (i.e how courts have applied copyright law in cases similar to one you may be interested in). Copyright only applies to "creative" works, where creativity is interpreted quite broadly.
In the case of a photo of a statue, copyright law does apply (the photo is considered a creative work), and copyright of the photo therefore belongs to the photographer (the photo being a seperate creative work than the scultpure itself). If multiple photographers took photos of it, each would own copyright to their own work.
However, in the case of a photo of a painting (vs statue), courts have previously found that copyright law does not apply since they've ruled that such a photo is a "slavish copy" lacking in any creativity. I doubt we've heard the end of such lawsuits though. One photo might be akin to a photocopy, but another might employ lighting, angle, composition, etc, and meet the (rather weak) standard of creativity that copyright law requires.
Software writing is a highly creative process, and most if not all software is going to be covered by copyright. If you just provided a software contractor with functional requirements specifications and left him to design and write the code, then you've not in any way removed the need for creativity (to do that you've had to have designed and implemented the code yourself!), so copyright law would indeed apply and as always the author of the creative work (programmer) by default owns the copyright.
If you wanted to contract for a piece of software to be written and you wanted to be granted certain distribution rights to it, then you'd need to specify exactly what rights it is that you are purchasing in the purchase contract. By default (if you don't put anything explicit in writing) then the programmer retains all rights. You may not like it, but that's the law.
If you feel that your requirements specification contains some valuable intellectual property that you'd like to protect, then you'd have to look into appropriate mechanisms to do that (trade secret, patent if applicable, etc), but that's entirely unrelated to who owns the copyright to a creative work based on those specifications.
If I paid someone to paint a painting for me, I own the painting.
Sort of... you own the physical painting, but that isn't the same as owning the copyright. Whether you owned the copyright (in whole or specific rights of duplication, etc) would depend on the purchase contract you arranged with the painter.
Same with a book. If you buy a book you own the physical object, but you don't own the copyright. You can't go making copies of the book and selling them yourself.
Same with software. If you "buy" a piece of software, what you've bought depends on the contract. You're typically going to be buying the physical media, and rights to use it either indefinitely or maybe for a limited period of time (cf renting a video), but not any copyright ownership unless that was specified in the purchase contract.
It doesn't make any difference whether you're buying an off-the shelf painting or software title, or comissioning someone to create a custom work for you. What you end up owning and the rights you have depend on the contract. Copyright law says that the author by default owns the copyright, so if you want anything different you need to put it in the contract.
There's no difference between paintings, software, books or any other type of copyrightable material - the same copyright laws cover them all.
I doubt it'd make any difference.
Speech recognition technology is really still in its infancy... it's possible to get good results but only under the most controlled of circumstances... high quality microphone, no background noise, clear diction, recognition engine trained for the speaker, etc. Even then it may depend on what you're actually saying, since in the case of any ambiguity a smart recognition engine will fall back to grammatical analysis and word frequency counts etc to try to guess right.
The real problem is that speech recognition requires artificial intelligence to do right and we don't have it. Often we understand speech that word for word is basically unintelligible, but we automatically apply context and intelligence to figure out what the speaker was trying to enunciate. Without full AI, a computer can't do that - it's much more limited.
When the foreigner serving you at McDonalds mumbles something after taking your order, you only need to understand a single word, or maybe not even that, to realize they are saying "to stay or take out?", but a computer today would need the person to have spoken clearly enough to have made out the words. Ditto for a stream-of-consciousness rambling GMail voice message with highway noise in the background - YOU may be able to FIGURE OUT what is being said, but that's not the same as the words actually being intelligible which is what a computer, without AI, would need to be able to transcribe it accurately.
Having a clear speech signal in the first place, or having a broadband vs telephone limited one, isn't going to make much difference, except under otherwise very controlled conditions.
No - look at the data from top of TFA (those two tables). IE was in 5th place 50% of the time vs what should have been 20% if it had been random..
It was a large statistically meaningful sample. The problem here isn't round-off or a poor random number generator. It's using an invalid list randomize algorithm.
Getting random numbers evenly distributed over an arbitrary range isn't a problem as long as you've got them available evenly distributed over *some* range. You just scale them.
No - look at the example of the two-element list in the link I provided. I'll copy it here for convenience:
A commonly used shuffle algorithm attaches random tags to elements
of a sequence and then sorts the sequence by the tags. Although this
algorithm performs well in practice, it does not achieve the perfect
random shuffling.
Let us consider the simplest example (which happens to be the worst
case): a sequence of two elements, [a b]. According to the
shuffle-by-random-keys algorithm, we pick two binary random numbers,
and associate the first number with the 'a' element, and the second
number with the 'b' element. The two numbers are the tags (keys) for
the elements of the sequence. We then sort the keyed sequence in the
ascending order of the keys. We assume a stable sort algorithm. There
are only 4 possible combinations of two binary random
numbers. Therefore, there are only four possible tagged sequences:
As we can see, after the sorting, the sequence [a, b] is three times
more likely than the sequence [b, a]. That can't be a perfect shuffle.
**
The example is assuming that of the two random numbers, a b - not a == b.
Never mind the theory, look at the results from TFA. IE was in 5th place (out of 5) 50% of the time in a large statistically meaningful result set.
Did you actually read TFA - at least the raw data at the TOP of the article ?!
Microsoft IE was in 5th place (out of 5) 50%, as opposed to a random 20%, of the time !!
Hard to say whether that's in their favor or not ... last place is just as special a position as 1st place.
Well, as it happens (if you read TFA) Microsoft's solution ends up being biased against Microsoft rather than for them. It hard to say which is worse, though; in one case you work against your own market share, in the other you end up wasting time and money as the EU hauls you back in court for unfair practices.
It seems the least cost option for Microsoft would have been to have gotten this right first time (as is always the case with software).
It's not an issue of how to get a truly random number, or of seeding a random number generator, but rather how to you use a source of random numbers to randomly order a list.
Some "reasonable sounding" methods don't actually work - e.g. attach random numbers to each list item and sort the list by these numbers (1). Microsoft used a similar method of sorting using a random comparator.
Some simple methods that DO work are picking a random permutation or executing a bunch of random swap operations on the list.
(1) http://okmij.org/ftp/Haskell/perfect-shuffle.txt
Hard to say whether that's better or worse than the common NASA "cost plus" contract where they agree to pay development costs plus an agreed profit margin. Where's the incentive for the developer to keep cost under control?
They've already done it:
http://mythbustersresults.com/episode50
I worked at Acorn in the early 80's and knew Steve Furber, but it's amusing to think that most of the Slashdot crowd probably wasn't even born then. I assume the average age here nowadays must be college age or thereabouts.
Not surprising they don't know what ARM originally stood for.
Build up a seawall an inch a year eh? And end up with a large part of the coastal area below sea level but protected by the seawall?
Kind of brings to mind a certain US city that was below water level protected by a retaining wall.
That "inch a year" of protection is all very well until you get a storm, surge or worse.
If NeXT sold to Apple for $400M, why did NeXT die?
It doesn't make much difference from the point of view of success, which is what we're discussing. Bottom line is he started a company and sold it it for $400M.
Presumably you know as well as me why NeXT (the company vs the software) died... I'd put it down to the difficulty of trying to get momentum for a brand new operating system.
I've sold software that I wrote for $40m. It's still in use. http://us.lastminute.com/
I find that hard to believe, given the small-minded nature of your potshots, but if it's true then bully for you. You do realize that NeXT software is still in use too, right? (Apple Cocoa, Objective-C).
Maybe "one hit wonder" is inaccurate, but "cult leader" may be sufficient.
He has what it takes to motivate people to success. No need to be disparaging about it in this context.
I wouldn't put Apple in that category... Look at Steve Jobs' track record and it's pretty amazing...
Apple, NeXT (sold to Apple for $400M+), Pixar, Apple again (Jobs return and resuscitation of the company via iMac, iPod, iPhone...).
That's a quite of a string of major successes, even if there have been a few product failures at Apple along the way. I'm not sure that Jobs himself has had many start-up failures.
So, it's not right to call Apple/Jobs a one-hit wonder. Maybe he's just serially lucky (SOMEONE has to be at the extremes of the bell curve of any phenomemon), maybe not, but that's a different story.
It's not that failure is good for you (although of course you can learn from it), but rather that it's pretty much inevitable, so you better learn how to plan for and deal with it.
For example, the success rate for start-up companies is quite small (10% - I forget), so if you're going to try a start-up it's best not to commit yourself to such a degree that it hurts your ability to shake off the failure and try again.. and again..
There's an interesting book about the start-up experience of AutoDesk (the company that created AutoCAD) called "The AutoDesk File" by John Walker, that says the same thing. AutoDesk's founders never expected to start a CAD software company... but in the end that was the product idea that became successful. The general conclusion was keep trying and let marketplace success not preconceived ideas dictate your level of financial/etc commitment.
Yes, and no ...
There's plenty of reasons to be suspect of WikiPedia, not least that officially it doesn't even strive for the truth - just for verifiyability (basically a published source).
However, there have been studies done showing that WikiPedia articles are on average just as accurate as Encyclopedia Britannica ones - both have similar average numbers of errors per article.
http://news.cnet.com/2100-1038_3-5997332.html
A large percentage of older programmers are unable to learn a new programming model. For example: Object Oriented coding.
I hate to break it to ya, sunshine, but OO design isn't anything new, even if languages to make it easy are more so.
Want OO design and polymorhism in 'C' or assembler? Just use a struct with pointers to functions. Things like implicitly called constructors and destructors are just syntactic sugar.
Us old fucks aren't quite as stupid as you imagine us to be! ;-)
For simple projects that's true - where the application itself is straightforward and you just need people conversant in the tools you need to use to build it. That's not why you hire highly experienced people. No technology really takes more that 5 years to thoroughly master (although some complex domains may do).
Where you DO need experienced people is where the application is significantly complex or critical, where what the experience buys you is expertise in software architecture, and design. There are many obvious high-profile jobs like this - compiler design, radar systems, speech recognition, etc, but also many much more mundane ones... complex multi-threaded realtime software at the core of communications systems, industrial control systems, etc.
For these types of projects what you really need is years of experience in designing big complex systems. Being a hotshot in C++ or Java or the technology du jour (however many/few years it took you to get there) doesn't help at all if the scope and complexity of the project is just way over your head. This doesn't just apply to massive software projects - there are plently of complex one-man projects where someone without significant design experience is going to be struggling, making mistakes, and making slow progress (and ending up with a bunch of unmaintainable code that reflects the struggles) as opposed to a more senior person who'll design it right and in a fraction of the time because the job was well below his experience level rather than above it.
For hobby use, with no particular application in mind, I'd suggest learning Python. If you enjoy that and want to get more serious about programming, then learn Java, C# or C++. If you have in mind WHAT you'd like to program, then different advice may apply.
Python is well designed and scales up reasonably well (unlike, say, Perl, which does neither). It's more fun to start with an interpreted language due to the instant feedback you get.
With COBOL you may have a point - OTOH are there really any full-time COBOL programmers out there? I thought they all went back into retirement after coming out for the Y2K feeding frenzy!
OTOH, say I'm looking to hire someone to do Java development... I'd MUCH prefer a C++ programmer with 10 years experience over a fresh graduate who knows Java. The C++ programmer will learn Java few months, but the newbie will need 10 years to get 10 years worth of programming experience.
The difference in productivity between a programmer with a lot of experience and a newbie can be astounding... A seasoned professional who puts his years of experience to play in designing it right can do in a week what a newbie may struggle months to get working (and still end up with an unmaintainable poorly written mountain of code vs the clean solution of the pro).
It seems that functionally Clutter has been superceded by advances in Qt.. Qt supports OpenGL/ES/VG backends and has a new "declarative UI" for designing animated and custom UIs. In fact since the Nokia aquisition Qt seems to have all but forgotten about the desktop and most new features are squarely aimed at implementing fluid custom iPhone-like interfaces on smartphone/netbook targets.
Looking at that Fisherman's Wharf in Google maps some more, in general they are just presenting a menu of geo-tagged photos, but the big round Wharf sign itself is a special case.... For that they've grouped a bunch of photos of the same object and have determined how these map onto a street view photo of the same object.
The fact that no other user photos (in this scene at least) are mapped to street view makes it appear this was done as a special case. Maybe the user photos were hand tagged as being of the same object? In any case, this is just mapping to a master photo, not integrating into a 3-D scene.
The user photos are not overlaid on the street view (neither were they in the Microsoft demo) - they are just presented as an alternative view of the same location. i.e. Google is just using geo-tags to give to a user-selectable bunch of photos taken at/near that location. Yawn.
That is completely unrelated to what Microsoft's demo did. It didn't just give you a menu of photos geo-tagged to the location, or simply overlay them at that spot. It actually recognized the content of the photos, where they belong in the 3-D scene, and incorporated them into the 3-D scene in augmented reality fashion. Way cool!
Huh?
You linked to a Google earth page that:
1) Let's the user drape a photo over the terrain layer (i.e. a pre-existing 3-D model)
2) Let's the user play "pin the tail on the donkey" with their photos
What do either of these have to do with 3-D model extraction from photos and mapping of photos into 3-D scenes?!!
What to either of these have even remotely to do with image recognition of any sort?!!
Clue: Those are rhetorical questions.