"exceptions are (in almost all implementations) much slower"
In Python, it can be much faster to catch SOMEWHAT rare conditions in exceptions than by using if tests to test for those conditions. So, the maxim "exceptions are for exceptional conditions" isn't considered paramount in Python as it is for, say, C++. Exceptions are becoming something more people are using for flow control when it's convenient to do so.
Ever since my high school math teacher showed this equation to me, I've had a problem with it because it is NOT elegant. It is artificially constructed to have both 1 and 0, when in fact it could be shortened to leave out 1 and have -1 on the right.
THAT would be elegant, but it wouldn't be able to lay claim to having so many important numbers in it.
It's a typical engineering case of a compromise between different aims. A good compromise, and an amazing formula, but still a compromise.
As a point of fact you are incorrect from a legal perspective.
Public use is prior art whether it's published or not. (There is, I'm told, controversy in the case that the invention was kept as a trade secret, but that didn't happen with 212-ROMANCE; not going to the trouble of publishing an article on something is not, legally, the same thing is making it a trade secret. Trade secret status requires actively keeping something secret. That did not happen; I talked about it quite openly without requiring NDA's.)
You say the purpose of the patent system is "to get people to document and publish their inventions" but that is only part of the purpose. The larger purpose is that it takes an expenditure of time and often financial resources to come up with a non-trivial invention, and if someone can subsequently copy it with impunity, there is less motivation to make that investment. So the larger purpose is to benefit the common good by providing motivation for that risk taking.
Now, if we're talking about an invention that is obvious enough that people have already created it and used it publicly, it's hard to also argue that it takes such an investment to create that particular invention that it wouldn't be created without patent protection. So the primary reason for patent law just doesn't apply to such a case.
When the powers-that-be weighed these different factors, what came out is that public use is prior art (with the aforementioned possible carve-out for a trade secret). You may disagree with the powers-that-be for making things that way, but I personally think they made a good compromise.
"Well, then, if his work in the 1980s counts as published prior art for the Pinpoint patent, then they also invalidate his own patents from 1998 and 1999."
That incorrect, because my later patents are not broad patents. They are very narrow covering specific mathematical approaches to solving problem, so they consitute an advance over my 1980's prior art. If, instead, my later patents attempted to be broad, my 1980's prior art would indeed invalidate them.
" Is the data distributed the same way? Does it represent a random process or not? Do users generate the same number of selections in the domains? Are there orders of magnitude more objects to recommend in one domain than in the other? "
The 1980 technologies were nearest-neighbor technologies. They provide prior art to somebody doing recommendations by nearest neighbors, no matter what the domain, I believe, because they will work, if not optimally, in domains with different distributions (frankly they are not even smart enough to care about the distributions) and different numbers of selections, etc.
But they are ONLY RELEVANT to NEAREST-NEIGHBOR-based CF.
I don't know what your stuff did, so I don't know if they're relevant to that. I see from your bio that you did the "customers who shopped for this item also shopped for..." stuff, which, by the way, is the only CF technology on Amazon I use -- and I use it a lot. I don't know if you use nearest neighbor stuff in it.
However, Pinpoint is suing Amazon, and I notice nearest neighbor verbiage in their claims. That's why I responded with 212-ROMANCE and Hey as prior art. But it isn't necessarily so since we don't know exactly what basis they are suing on -- there are a LOT of claims there. I should update my piece to reflect that this 1980's prior art may not be relevant because of the fact that it is nearest-neighbor based.
I feel moved to say more on those patents. I really thought they might have value if anybody wanted to take advantage of my long and hard work in developing certain arcane mathematical techniques. But I made NO effort to make them readable, presentable, etc. I had NO time to give it. My lawyer put some meaningless gobbledegook in the abstract and in the intro, etc., and I just didn't care what it said because it had no legal bearing I had other things to do. It was just nonsense as far as I was concerned.
I certainly didn't expect back then that people would be reading the abstract in isolation in a place like Slashdot -- I couldn't have dreamed it.
So it wounds like it may be YOUR work in particular that is coming under attack?
Can you see any way Pinpoint can claim to have come up IP that covers your specific technology?
Responding to the point you made above:
The technologies being discussed in my piece as prior art, the 212-ROMANCE one and the Hey one, could be used in those different domains. They would not be the best possible ones in those domains -- in particular 212-ROMANCE did have a very robust mathematical basis -- but they would WORK if you didn't have anything better. They are very general-purpose strategies. They are far from the best possible strategies, but that's not the point when the goal is to contest an overly-broad patent which is trying to cover all possible strategies.
It's really hard for me to see how they think they could use the Pinpoint patent against Amazon if they aren't doing exactly that.
The abstract isn't meaningful legally. If you look at what is actually claimed and read the specification to see what it really means, you'll see that it does not have the admittedly crappy overly broad meaning that you took from the abstract. But the abstract just doesn't matter. This is an extremely narrow, mathematically-oriented patent.
Er, you did notice that from the moment this story was originally posted, it contained: " Full disclosure: I have a couple of patents of my own in the CF space. However, these patents are for highly specific mathematical techniques and are of no interest to anyone using different mathematical techniques, of which many are freely available."
Right?
And they aren't "backup" to the 8-inch media since they were filed in 1997 -- too late to do anything with respect to the 1994 root of the Pinpoint stream of continuations.
One other aspect, since we're discussing these legal technicalities... if you create an actual reduction to practice, and then use the invention publicly, you have one year to file a patent application or your own public use becomes prior art. So if you want to wait years, you have to keep it totally under wraps, but have solid proof of the relevant dates.
You're incorrect. The diligence is required between the date of conception and the actual reduction to practice. After that, you are merely required not to "abandon" the invention. Abandonment does not occur if you even have an intention of working on it again.
Yeah, actually I copied it to 3.5-inch mac floppies at some point as well as to DAT. And I probably have it somewhere on my hard disk on the laptop I'm typing this from as well.
A clarification, it appears that it would apply to text ads only if there was a limited amount of space on the page, so that bidding was used to determine which ads would appear.
It appears to be about the situation where people bid for ad space. Emergent Music does that. A lot of sites do, it arguably includes the phenomenon of "text ads" such as are used on Kuro5hin.
One thing that is important to keep in mind. This isn't a patent, but a published application. The patent has not been issued or approved by the PTO.
However, once a patent is published, companies that would infringe if it were issued can be contacted, and if they continue to use the claimed matter, can be sued for damages retroactively when it is issued.
The patent was applied for on October 8, 2002. It seems obvious that there was plenty of prior art. I don't know, though, whether the patent "takes priority" from some earlier application such as a provisional patent application. I'm not sure whether the online publication mechanism normally makes that information available.
As I get more info I'll post it to my blog.
Hi just thought I'd chime in, I'm CEO of the company that put Emergent Music together, and I definitely agree it's a good place to go to get your music out there.
Slashdot Mirror
"exceptions are (in almost all implementations) much slower"
In Python, it can be much faster to catch SOMEWHAT rare conditions in exceptions than by using if tests to test for those conditions. So, the maxim "exceptions are for exceptional conditions" isn't considered paramount in Python as it is for, say, C++. Exceptions are becoming something more people are using for flow control when it's convenient to do so.
Ever since my high school math teacher showed this equation to me, I've had a problem with it because it is NOT elegant. It is artificially constructed to have both 1 and 0, when in fact it could be shortened to leave out 1 and have -1 on the right. THAT would be elegant, but it wouldn't be able to lay claim to having so many important numbers in it. It's a typical engineering case of a compromise between different aims. A good compromise, and an amazing formula, but still a compromise.
As a point of fact you are incorrect from a legal perspective. Public use is prior art whether it's published or not. (There is, I'm told, controversy in the case that the invention was kept as a trade secret, but that didn't happen with 212-ROMANCE; not going to the trouble of publishing an article on something is not, legally, the same thing is making it a trade secret. Trade secret status requires actively keeping something secret. That did not happen; I talked about it quite openly without requiring NDA's.) You say the purpose of the patent system is "to get people to document and publish their inventions" but that is only part of the purpose. The larger purpose is that it takes an expenditure of time and often financial resources to come up with a non-trivial invention, and if someone can subsequently copy it with impunity, there is less motivation to make that investment. So the larger purpose is to benefit the common good by providing motivation for that risk taking. Now, if we're talking about an invention that is obvious enough that people have already created it and used it publicly, it's hard to also argue that it takes such an investment to create that particular invention that it wouldn't be created without patent protection. So the primary reason for patent law just doesn't apply to such a case. When the powers-that-be weighed these different factors, what came out is that public use is prior art (with the aforementioned possible carve-out for a trade secret). You may disagree with the powers-that-be for making things that way, but I personally think they made a good compromise.
"Well, then, if his work in the 1980s counts as published prior art for the Pinpoint patent, then they also invalidate his own patents from 1998 and 1999." That incorrect, because my later patents are not broad patents. They are very narrow covering specific mathematical approaches to solving problem, so they consitute an advance over my 1980's prior art. If, instead, my later patents attempted to be broad, my 1980's prior art would indeed invalidate them.
" Is the data distributed the same way? Does it represent a random process or not? Do users generate the same number of selections in the domains? Are there orders of magnitude more objects to recommend in one domain than in the other? " The 1980 technologies were nearest-neighbor technologies. They provide prior art to somebody doing recommendations by nearest neighbors, no matter what the domain, I believe, because they will work, if not optimally, in domains with different distributions (frankly they are not even smart enough to care about the distributions) and different numbers of selections, etc. But they are ONLY RELEVANT to NEAREST-NEIGHBOR-based CF. I don't know what your stuff did, so I don't know if they're relevant to that. I see from your bio that you did the "customers who shopped for this item also shopped for ..." stuff, which, by the way, is the only CF technology on Amazon I use -- and I use it a lot. I don't know if you use nearest neighbor stuff in it.
However, Pinpoint is suing Amazon, and I notice nearest neighbor verbiage in their claims. That's why I responded with 212-ROMANCE and Hey as prior art. But it isn't necessarily so since we don't know exactly what basis they are suing on -- there are a LOT of claims there. I should update my piece to reflect that this 1980's prior art may not be relevant because of the fact that it is nearest-neighbor based.
I feel moved to say more on those patents. I really thought they might have value if anybody wanted to take advantage of my long and hard work in developing certain arcane mathematical techniques. But I made NO effort to make them readable, presentable, etc. I had NO time to give it. My lawyer put some meaningless gobbledegook in the abstract and in the intro, etc., and I just didn't care what it said because it had no legal bearing I had other things to do. It was just nonsense as far as I was concerned. I certainly didn't expect back then that people would be reading the abstract in isolation in a place like Slashdot -- I couldn't have dreamed it.
So it wounds like it may be YOUR work in particular that is coming under attack? Can you see any way Pinpoint can claim to have come up IP that covers your specific technology? Responding to the point you made above: The technologies being discussed in my piece as prior art, the 212-ROMANCE one and the Hey one, could be used in those different domains. They would not be the best possible ones in those domains -- in particular 212-ROMANCE did have a very robust mathematical basis -- but they would WORK if you didn't have anything better. They are very general-purpose strategies. They are far from the best possible strategies, but that's not the point when the goal is to contest an overly-broad patent which is trying to cover all possible strategies. It's really hard for me to see how they think they could use the Pinpoint patent against Amazon if they aren't doing exactly that.
The abstract isn't meaningful legally. If you look at what is actually claimed and read the specification to see what it really means, you'll see that it does not have the admittedly crappy overly broad meaning that you took from the abstract. But the abstract just doesn't matter. This is an extremely narrow, mathematically-oriented patent.
Er, you did notice that from the moment this story was originally posted, it contained: " Full disclosure: I have a couple of patents of my own in the CF space. However, these patents are for highly specific mathematical techniques and are of no interest to anyone using different mathematical techniques, of which many are freely available." Right? And they aren't "backup" to the 8-inch media since they were filed in 1997 -- too late to do anything with respect to the 1994 root of the Pinpoint stream of continuations.
One other aspect, since we're discussing these legal technicalities... if you create an actual reduction to practice, and then use the invention publicly, you have one year to file a patent application or your own public use becomes prior art. So if you want to wait years, you have to keep it totally under wraps, but have solid proof of the relevant dates.
Phew, I feel vindicated now. :)
You're incorrect. The diligence is required between the date of conception and the actual reduction to practice. After that, you are merely required not to "abandon" the invention. Abandonment does not occur if you even have an intention of working on it again.
And don't forget I have the hardcopy.
Yeah, actually I copied it to 3.5-inch mac floppies at some point as well as to DAT. And I probably have it somewhere on my hard disk on the laptop I'm typing this from as well.
Thanks, I'll fix that. :-p
A clarification, it appears that it would apply to text ads only if there was a limited amount of space on the page, so that bidding was used to determine which ads would appear.
It appears to be about the situation where people bid for ad space. Emergent Music does that. A lot of sites do, it arguably includes the phenomenon of "text ads" such as are used on Kuro5hin. One thing that is important to keep in mind. This isn't a patent, but a published application. The patent has not been issued or approved by the PTO. However, once a patent is published, companies that would infringe if it were issued can be contacted, and if they continue to use the claimed matter, can be sued for damages retroactively when it is issued. The patent was applied for on October 8, 2002. It seems obvious that there was plenty of prior art. I don't know, though, whether the patent "takes priority" from some earlier application such as a provisional patent application. I'm not sure whether the online publication mechanism normally makes that information available. As I get more info I'll post it to my blog.
Hi just thought I'd chime in, I'm CEO of the company that put Emergent Music together, and I definitely agree it's a good place to go to get your music out there.