What is the fastest sustainable web service response time?
in [MFC]
|
From: Hector Santos on 24 Feb 2010 10:45 Peter Olcott wrote: > "Hector Santos" <sant9442(a)nospam.gmail.com> wrote in message >>>> GO AWAY PATENT TROLL! >>>> >>> (1) You did not use the term "patent troll" correctly, >>> this term only applies to entities that have no intention >>> of commercialization. >> >> No, a Patent troll is one that frivolously patents the >> idea that the shortest distance between two points in a >> straight line. > > Wrong! > http://en.wikipedia.org/wiki/Patent_troll > Words and phrases do not mean whatever you want them to > mean, they only have meanings that are commonly agreed upon > otherwise communication would be impossible. Patent troll is currently a controversial term, susceptible to numerous definitions, none of which are considered satisfactory from the perspective of understanding how patent trolls should be treated in law. [1] That said, you are a PATENT TROLL: - who takes pre-existing ideas with NO new non-obvious contribution, combines them to claim patent D. Unfortunately that is what current flawed patent guidelines has allowed since 1996. You obviously have no understanding on this concept (because you are a PATENT TROLL) and how it can be apply and in most cases, does not apply to preexisting implementators who do not or did not have the relaxations in patent laws to allow patentability. So even if there was NO patents, attempting to enforce your frivolous patent may not apply! - who patrols the technical forums, ask novice questions, gets answers and tries to get IP protection simply because it isn't in the books. And then has the cojones to tell people "Its my technology." - This is NEW WORST FORM OF PATENT TROLLS - YOU! > All that it takes to create something new under the Sun > (according to the patent office) is to combine two or more > pre-existing concepts to provide a substantial new benefit. > As long as this combination is not shown in published prior > art, it is patentable. In some cases it is patentable even > if the combination is shown in prior art, if the new use has > not also been shown. Classic mindset of a PATENT TROLL! Not a true inventor of new ideas but a leach of existing ideas. > I think that 20,000 hours worth of effort deserves some > financial remuneration. Try suing some companies and see how far you will get. This is my last note on this. You can have the last word - PATENT TROLL! -- HLS [1] http://en.wikipedia.org/wiki/Patent_troll
From: Mihai N. on 25 Feb 2010 05:32 > As long as this combination is not shown in published prior > art, it is patentable. In some cases it is patentable even > if the combination is shown in prior art, if the new use has > not also been shown. "Patentable under the current law" is not the same as "worth a patent" and "legaly ok" is not the same as "morally ok" Also, a requirement is to be non-obvious "to one of ordinary skill in the art." To take a font, render the alphabet with that font at a certain size, then compare the resulting bitmap with a bitmap received from somewhere else is the most obvious (and naive) approach to OCR. ("naive" because it does not work too well) This is not the aproach taken by currnt OCR solutions because it was tried and does not work, not because is so creative that nobody thought about it. > I think that 20,000 hours worth of effort deserves some > financial remuneration. Even 8 hours deserve financial remuneration. Hell, even 10 minutes! But that should be based on the fact that someone on the open market is willing to pay for a product that brings some value. Not because of some patent that gets in his way of crating something. If I remeber well, we've been thru this before (in 1996, if I am not wrong), when I have tried to explain how what you want is not going to work too well. Now it looks like after 4 years you still have no product (visiting www.SeeScreen.com) and you are trying to maybe extort someone by dropping a patent in the middle of the road. -- Mihai Nita
From: Mihai N. on 26 Feb 2010 01:30 > If you think it is not worth a patent, you need to change the law. Oh, how much I would like to do that! Meantime, it is not morraly ok to abuse a broken law. If the law in country X does not protect kids from exploitation, it still not moraly ok to use those kids for cheap labor. And maybe it not even the law, it is the system implementing it. Requirement for the patent to be non-obvious is a good thing. The fact that "obviousness for someone in the field" is evaluated for people who are not practitioners in the field is wrong. > If it had been obvious, it would have already been done. I have a problem with that statement. Many things have been done before, but not *exactly* the same. The "wheel" in iPod was around since forever (think the old radios with a wheel to change the scale, around since WWII, or professional VCRs using the wheel to control tape advance). Oh, but that was a radio, the "new creative thing" it not playing music from the waves, but from a digital file. And I am not changing the scale, but advancing to the next song. How creative! How unique! So worth a patent! My small tool that scans C/C++ code and replaces ansi calls with generic calls is obvious and trivial. But nobody did it before (at least not for internationalization APIs). Is that a good argument for patentability? No! So no, in my book "was not done before" is not a good argument. If some brute-force aproach was not done because the computing power was not there, but 20 years later it can be done, does it suddently make it patentable, because "it would have already been done"? Is it ok to patent ideas that are not possible today? Like: patent and method of instant package delivery system. I patent, wait for someone else to invent teleportation, profit :-) -- Mihai Nita [Microsoft MVP, Visual C++] http://www.mihai-nita.net ------------------------------------------ Replace _year_ with _ to get the real email
From: Jerry Coffin on 7 Mar 2010 17:03 In article <6ocdo51sb5ol3uvat85s57nihoolru6na9(a)4ax.com>, newcomer(a)flounder.com says... [ ... ] > understand the purpose of the patent law: it is to reward > innovation by giving the innovator exclusive rights to the patented > object, design, or methodology for a fixed number of years (17 > years for general patents; different times for patents on > automotive technology, designs, and a few other exceptions, all of > which are shorter). Nothing more than that. Well, actually, the purpose is quite a bit more than that. It's not just a one-sided reward for innovation -- it's an exchange. Specifically, the patent is required to contain teaching of how to use the patented invention, so when the patent expires the public has a right to use it, AND an instruction manual (so to speak) telling what to use it for ("best mode") and how to implement it. [ ... ] > But the law clearly states the prerequisites for a patent, and to > knowingly submit a patent when there is a failure to meet the prerequisites is immoral. The ramifications are not only moral, but legal -- though it's difficult to do so, there's at least theoretically the possibility of invalidating a patent based on "unclean hands". > There is no "morality" to a patent. It is either valid, or invalid. That's open to argument -- there are enough different ideas of "morality" in the world that I don't think anybody can definitively state that "there is no morality to X" for *any* value of X. [ ... ] > Consider also whether the RSA patent on Public Key encryption is > valid. Whitfield Diffie laid out the principles of PKI years ago > (I went to one of his talks at CMU when I was there) but RSA is a > *particular realization* of the principles of PKI). More tellingly, the British intelligence service had invented exactly the realization of PKI used in RSA years before -- but failed to grasp its implications. They saw it only as a mathematical curiosity, and abandoned it. Whitfield Diffie came at it from the opposite direction: knowing what he wanted and what it could accomplish long before anything had been published about how to do it. Years later, when Diffie-Hellman (and later still RSA) were invented, he'd already plowed the field, so to speak, and it grew quickly. > On the whole, I am not a fan of software patents because I think too many of them are > granted erroneously for ideas that are not only screamingly obvious to a practitioner in > the field, but many of which had prior art implementations unknown to the patent examiner. The biggest problem, at least IMO, is that for quite a while the PTO routinely rejected _all_ patent applications that involved software at all. Though you and I know other publications are legitimate prior art, when the PTO searches for prior art, nearly all the usually look for is prior patents/applications. Since there are essentially none up until the '80s when they were forced to start accepting software patents, they lack a record of a great deal of prior art, without which they have no basis to reject applications for patents on methods that have been fairly well known for decades. Worse, most of those older records only exist in paper form, so searching them electronically can be quite difficult. Just for example, a few years I was involved in a patent case where we invalidated a patent based on a PhD dissertation -- which was only available in paper form at the college where it was written (in Australia, no less). Fortunately, it was cited in a newer paper that was available electronically, and we chased it down from there... > Ask someone how to do OCR. Ask if the simplistic pattern-matching algorithm is viable. > The answer will be "no". So the profession "teaches away from" the methodology Peter is > using. He came up with some new ideas of organizing the computations in a way that makes > them computationally feasible. So it was *not* obvious that this should work (and what > *was* obvious was that it was the wrong approach). In fact, that's one of the tests the courts have formulated for rebutting a rejection based on obviousness -- that the prior art has "taught away from" what is being patented. (e.g. in re Geisler and in re Woodruff). [ ... ] > I'm also reacting to the attitude that software patents are > inherently immoral, and that anyone who applies for one is > therefore tainted with the troll-brush. I do not think the > examples of poorly-researched prior art (one of the serious > weaknesses of the patent system) necessarily make all software > patents invalid. I think there's a bit more to the attitude than that. When I read diatribes against software patents, what comes most strongly to mind is the stories of the old west, with settlers trying to build fences, and cowboys tearing them down. Patents fulfill the same basic purpose as fences (setting boundaries on what particular people possess), and many of the current software practitioners have pretty much the "cowboy" attitude -- that no law should apply to them, and placing any limits on what they can do or where they can go is inherently wrong. > And as to the appropriateness of software patents, go read Pam Samuelson's articles on the > subject. As with AOL compliance issues, "you may not like the result, but it the law. And > the only recourse you have if you don't like the law is to change it". Sadly, the firms > that would benefit most from a sui generis law on software are afraid to rock the boat and > lobby for a new law, thinking they will be worse off than they are now (Pam thinks > otherwise). There has been a movement toward patent reform lately, but little of it is of the sort that your typical programmer would appreciate. Personally, I think some pretty minor changes could make big differences. The single biggest problem I see is that right now, the level of proof required to attain a patent is only "preponderance of the evidence", but invalidating that patent later requires "clear and convincing" evidence. Simply changing that to use the same standard for both would fix (in my estimation) about 80-90% of the biggest problems. -- Later, Jerry.
From: Hector Santos on 7 Mar 2010 17:47
Jerry Coffin wrote: >> On the whole, I am not a fan of software patents because I think too many of them are >> granted erroneously for ideas that are not only screamingly obvious to a practitioner in >> the field, but many of which had prior art implementations unknown to the patent examiner. > > The biggest problem, at least IMO, is that for quite a while the PTO > routinely rejected _all_ patent applications that involved software > at all. Right, many in the legal circle 1996 (guidelines relax) as the now "Timeline" "Clock" where prior art meant less or rather may be deferred in footnotes. > Though you and I know other publications are legitimate prior > art, when the PTO searches for prior art, nearly all the usually look > for is prior patents/applications. Exactly. That is the problem. However .... > Since there are essentially none > up until the '80s when they were forced to start accepting software > patents, they lack a record of a great deal of prior art, without > which they have no basis to reject applications for patents on > methods that have been fairly well known for decades. Being in the telecommunications business, specifically the Online Hosting, BBS market since the 80s, I have seen over the years on going claims on basic features that was obvious and common on market. I just shake my head. One reason we were not worry because the USPTO does recognize existing implementators can not be denied to exist or even naturally evolve with an obvious new feature that is natural part of your process or product. A few years back, one a-hole patented how to handle a SMTP response code for challenge-response systems. His premise that SMTP systems do not use the "RETRY" (45x) response code thus he frivolously patented the USE of 45x. Regardless if his patent is valid or not, the courts will recognize existing systems can not be DENIED to use a 45x if they decided to do so. Another a-hole tried to patent the idea of Callback Verifications via the WEB. When I found out (by reading a "Patent Pending" claim at his web site), I wrote to him and provided a link to our Support File Library with a report of at least 150+ 3rd party written, Operator written CBV for dialup, telnet and the web from the 80s, 90s and 2000s. His response? "Thanks, I will update our patent application to add it to the prior art reference" In the past (prior to 1996), he was subject to a fines and criminal charges for intentional neglect of prior art. > Worse, most of those older records only exist in paper form, so > searching them electronically can be quite difficult. Exactly. I guess this is the reason the USPTO began to provide leeway in prior art references. But once found, its not to reject a patent, but to update it - document it. What it basically means is that the patentee can only go after those who did not have prior art implementations. > Just for example, a few years I was involved in a patent case where > we invalidated a patent based on a PhD dissertation -- which was only > available in paper form at the college where it was written (in > Australia, no less). Fortunately, it was cited in a newer paper that > was available electronically, and we chased it down from there... The other problem is that with the 1996 changes, the USPTO lost experienced Patent examiners because prior art was less important, and the dollars was better outside the government too. They began to hire younger people who simply took things for granted. You really do need patent examiners that have 20 to 30 years experience. They should be aggressively recruited retired engineers. > Personally, I think some pretty minor changes could make big > differences. The single biggest problem I see is that right now, the > level of proof required to attain a patent is only "preponderance of > the evidence", but invalidating that patent later requires "clear and > convincing" evidence. Simply changing that to use the same standard > for both would fix (in my estimation) about 80-90% of the biggest > problems. I agree, not sure if 80-90%, but a good percentage. But practically speaking, it boils down to people, and simply getting patent examiners with broad industry experience would resolve another large percentage as well. This was a recognized problem and I see today the USPTO is hiring!! Also, there should be a statue of limitation. Once granted a patent, if YOU knew a company, individuals or group were using the ideas in some form or fashion and if you don't approach them because they didn't threaten you within X years, you can not approach them LATER when they do capture a larger market share. Many patent holders wait until its viable to sue you, i.e. its been known IBM don't care about you until you hit 2% of their market. Take last week's APPLE vs HTC lawsuit. This is one interesting lawsuit that is going to have many repercussions (or should). IMV, HTC is just a proxy to the better picture. Apple is so scared here, they used 20 patents in the lawsuit. To me, the 1996 Patent #5,481,721, which hits home, is the big one. To me, if this patent prevails to weigh high in the lawsuit outcome, it will give APPLE complete control in the blossoming client/server GUI client small device telecommunications market for the next 4 years (patent expires in 2014). -- HLS |