Slashdot Mirror
Google (Patent Pending)
Jason Eric Pierce writes "I just noticed something interesting while using Google, considering how much it has been praised here by those that tend to denounce software patents:
'Unlike other current generation search engines, which rely heavily on keywords or meta-search technology to find information, Google offers an advanced patent-pending technology called PageRank(TM) to deliver the most relevant results. PageRank ensures that the most important, relevant pages always come up first and that your users will always find what they are looking for.' I love Google, but I hate the idea of patenting a search algorithm/implementation." Good point. Thoughts?
http://www.google.com/search?q= more+evil+than+satan and see the first hit?
Although software itself is not patentable (and should not be), technologies and algorithms definitely are. The distiction is (and should be) made by the patent office. Self-discretion in this context just means lack of money ;-)
I dunno. I agree that the vast majority software patenting that's been going on lately is silly; "One click ordering" is a pretty obvious thing.
But from what I've read, Google's search logic is pretty ingenious, and not very obvious at all.
Is this not an example of a case where a patent makes sense? At the moment I'd say that I am not troubled much at all by this one.
Patent #93,593,376:
A method of testing web servers by writing an interesting story for nerds, adding a few large JPG's, and then posting said link to slashdot.org. The result of posting is a large and immediate increase in web traffic, not for purposes of advertising (because said nerds don't have any money to buy things) but only to test the reliability of one's web server.
Licensing will be a no-brainer. If a company wants to test their servers, they can just mirror a six-month old news story and tell Hemos about it. "Psst - there's an article on here about why Linux web servers aren't as good as NT. It's only six months old - it's still news!"
- Sorry, Hemos, you walked right into that one.
What's your damage, Heather?
I don't think most Slashdotters are against all patents, but more against patenting the obvious (like the icon for these stories indicates). One-click shopping, or banner ads are so obvious (to one 'skilled in the art', eg. a geek) that they shouldn't be patentable.
What Google uses is far from obvious, so this merits a patent. I think this was the original idea behind the patent system, but somehow the USPTO seems to have lost the definition of 'obvious' along the way.
superblog.org: all your favourite blogs on o
This is the crux of the question is do slash.dotters object to the principle of patenting ideas full stop, or, just the patenting of dumb ideas.
If your priciple objection is to the patenting of dumb ideas then you should have no problem with the google patent. Thier search algorithms are certainly different from all the others and produce good and consitent results. It seems like they are patenting something "original, not obvious, and which works". There may be some prior art here as a ton of work has been done on search algorithms over the years, but generally speaking this looks like a "good" patent.
If the main objection is to patents per se, then I would say we are a bunch of hypocrytes. The whole high tech industry which produces all the goodies we love to play with is driven by patents. IBM is not going to spend billions researching "copper" etc. and give as those lovely gigahertz processors if some company in tiawan can rip off the design as soon as it is working.
Old COBOL programmers never die. They just code in C.
Search engine admins are constantly in a cat and mouse game with the Web. It's not just cat and mousing with word spammers (first rank down pages with words in alphabetical order... word spammers strike back... then try to parse pages a bit.. word spammers strike back... et cetera...). Styles of web page writing alter over time, and there are issues over the mutability of pages and the such. (Frames fall into fashion.. frames fall out of fashion..)
Search engine admins should have the freedom to tinker with their ranking algorithms without getting nailed by patent sharks. Besides, copying a search algorithm is a bit difficult if the algorithm isn't published. Search enginges don't have to publish their methods to build a user base. They just have to do a good job.
So, Google is probably justified in seeking a patent. One can't be too carefull. But if Google starts pulling an Amazon, it's time to boycott them.
I love Google, but I hate the idea of patenting a search algorithm/implementation
Well then how selfish are you? You love Google but you're opposed to them trying to make a buck from the technology for which you love them. You'd rather they didn't patent it, and someone like Lycos could come along and say "Thanks for doing all the research Google, this technology will work great in our new search engine".
And you know the masses will all go visit Lycos and click on their banner ads, because instead of wasting money on "research" and/or "development", they've been spending their money on ads on the sides of buses.
And then Google will go bust, and all their workers will go home and have no money to put food on the tables of their families.
But it's ok, because patents are bad, because Slashdot said so.
Anyone whose search engine returns this result for More Evil than Satan Himself can have a patent. That's one SMART search engine.
No Zen is good zen
There is a great deal of difference in someone owning a patent and wielding it when they shouldn't be. I think whilst we live in a world where there are software patents it is a good idea for google to make sure that they have the one for the technology they created, rather than let some ignorant person manage to give it to a competitor, who might use it against Google.
I know that's a bit unlikely, but do you really think Barnes and Noble were thinking 'AHA! we've pulled a swift one on Amazon - they won't realise we've stolen their technology? Of course not, they were just using an obvious idea - the fact that they didn't get the patent process started quick enough must be the only reason Amazon got the patent and not them.
So rather than let someone else try and get it, Google should get it themselves, but not be aggressive with it...
Since we're going to discuss PageRank and it's patentability, You might want to read about it first.
No Zen is good zen
the problem is when they're enforced where they shouldn't be... what amazon did was less protecting their rights, and more trying to look like Microsoft and beating up their competitors.
/. folk seem to put a lot of faith in them, I think they'd do the right thing if it came up.
If Google didn't patent this, whose to say Altavista wouldn't patent it tomorrow and sue them? Hopefully, if someone wants to make a search engine similiar to Google's, they'll be nice enough to keep from ripping them to pieces...
I don't really know, I don't know the Google folks too well. But a lot of
Devilled Eggs - A disturbing little creation of mine.
Which was actually the intent. The idea behind patents is that the person who thought of it should have a monopoly. The trade-off is that in exchange for that monopoly:
- The technology must be published, so that everybody can see it;
- The monopoly expires after a relatively limited time (in the U.S., 20 years). After that, it's fair game for anybody.
This 20 year time period seems a bit long in the Internet era (which works in dog years), but it's worked for a long time....phil
...phil
"For a list of the ways which technology has failed to improve our quality of life, press 3."
The original thinking behind patents, IMHO, is to encourage the development of novel products. Google has a superior product based on a search algorithm that is head and shoulders above the rest, both in originality and accuracy. In order to fully benefit from their superior product, they need to have their bread and butter insured against the actions of parasitic organizations that would otherwise leech ideas off of pioneering companies (in this case Google). Patents are not bad, they are just more susceptible to abuse than they could be. Fault lies with the USPTO, not the intrinsic concept of patents itself.
.{redmist}.
-------------------------------------------------
-------------------------------------------------
The relevancy of hits in Google seems to be pretty good, but the "I'm feeling lucky (TM)" thing IMHO is a dumb feature.
The fact that no-one else was doing this doesn't mean that no-one else thought of it or that it isn't obvious.
It would be nice if someone familiar with patent laws could tell us whether "it's obvious, but I'm the only one who believed in it" is grounds for a patent. I'm not sure how I feel about that myself, but I'm leaning toward "no".
--
Fuck the system? Nah, you might catch something.
I agree. Cached pages are a great feature. If someone were attempting to make this a vorbotten practice, I would be most unhappy. If one does not wish the public to view information, he should not commit to posting it on to the internet in the first place. Once such a transaction has been established, it is in public view. Some people wish to subject us to a bunch of silly laws for to further their own greed and cause problems for the rest of us. I say get rid of those silly people and make them publish books.
Speaking of ``the right thing'', wouldn't it be ingenious of them to pull an RSAREF on PageRank -- but rather than say ``for noncommercial use only'', require use of the algorithms to be GPL'd?
Kind of a silly idea, and certainly will not sit well with quite a few OSS folk (I personally find some fault in it), but it's a neat idea nonetheless.
The other day, someone was asking how long the internet boom could continue. I think the boom is part of a natural shift into an information based economy. Countries that restrict the flow of information by artifical means, be it censorship of the net in Australia or artificial restrictions imposed by patent and trademark laws which require anyone with a web presence to have a substantial and expensive legal team, will end up killing the boom in their borders. Said countries will end up being the new third world countries, strangling the wealth that the information age offers. Countries that make an effort to embrace the internet and assist the flow of information will most likely enjoy the boom indefinitely.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Google seems to have an aversion to advertising and keyword sales. Why are they bothering with a patent if they don't seem to be interested in generating a cent of revenue through normal means? Is it because of terms of some government funding they've gotten?
If Google's search engine is so original and they manage to keep it on tracks then there will be minimal problems. Maybe we will loose some dynamism in its progress but it is hard to level this point. Meanwhile if they do not touch similar algorithms that may slightly resemble their engine then we have all doors open to progress and an healthy competition.
However if they jump on the bandwagon on "how unique" their search engine is and stop making any serious developements on it. If, beyond this, they start playing lawsuits over similar systems. And if they simply try to hold market tighten to their original idea then we will have a lot of trouble. We have tons of examples on how such policies killed whole technologies. I cannot evaluate what benefits/losses we got from such thing as STAC vs Microsoft. But we can be sure that this story placed "disk compression" systems into a marginal alley. Today we have 6, 9, 13, 20, 28 Gb disks. Most of us may not feel how critical such systems can be sometimes, even if you have a lot of space laying around for most of the time. However, there are situations when a flexible, fast and light compression system would do a lot of help, specially in some extreme moments. However there are a lot of caveats on those systems that exist today. And I believe that this happens because disk compression is still healing from this conflict.
Let's give them the benefit of the doubt for a moment and assume that they have a legitimate non-obvious patent claim here. Then from a practical point of view, wouldn't it be great if we could persuade them to allow use of the patent in all free (speech) software projects. It would aid their PR, and would still protect them from people like Altavista. Note that due to the very nature of patents, this would enforce a form of copyleft on any software using their patent - making the software non-free would impinge the patent licensing agreement.
I might add that this is simply a practical compromise to the situation, and not a general solution for patents. If anyone here were ever to patent anything, then I would strongly recommend them to put the patent in the intelectual public domain. Patenting, unlike copyright, takes away more that just the right to someone's work - it takes away the right to someone's thoughts. Now whether you live in RMS's communist (from community) society or ESR's anarchist one, this impinges on your basic communal/individual freedoms and is hence a Bad Thing.
Still, until we can persuade the (US) government that software patents in their present form don't work -- and many of the upcoming court cases should hopefully help us do that -- approaching the company involved seems to be the only alternative to putting our head in the sand and hoping they don't see us.
it takes away the right to someone's thoughts
How the HELL do you reach this conclusion?
The only thing a patent grants the owner is a limited ability to prevent others from using an invention. Google comes up with a new search engine and patents it. Anyone can still think all they want about the methods used, and in fact do R&D on it without infringing. In fact, without patent protection the chance is you WOULDN'T be thinking about the Google algorithms because they would be kept as a trade secret.
Governments take away your rights every time they make a law. The laws against murder take away your rights to shoot somebody just because you have a headache.
The point of laws is that the OVERALL gain is supposed to outweigh the loss of freedom associated with having a law. Back in the 17th century governmants worked out the principle that granting patents in exchange for requiring the inventor to publish his technology was a benefit to society as a whole. If it weren't for patents, the Google inventors WOULD NOT be publishing their algorithms and you would have no idea how their search engine worked unless you tried to re-engineer it yourself, a needless duplication of effort.
until we can persuade the (US) government that software patents in their present form don't work
I won't argue with that - any system can be improved.
- Everybody knows about more evil.
- Interesting results are brought by crappy office suite, as well as by crappy os and crappy software in general.
- Guess where it'll land with bill the borg?
- But try bloated browser for a change.
Moderate this down (-1, Underwhelming)--
Industrial space for lease in Flatlandia.
Well, perhaps their enthusiasm for free/open source software (unless paid mercenery astroturfers count as slashdotters these days), but beyond that I doubt you'll find slashdotters agreeing on much of anything.
Some of us (such as myself) believe that patents should be eliminated entirely, that they are a diservice to humankind and do more to harm and slow down technological progress in all areas of scientific endeavor than any other single thing.
Others are against software patents, but do not feel the same reasoning applies to other disciplines.
Others are simply against the pathetically obvious patents being issued by the USPO, and would like to replace the people issuing these patents but keep the system as it is largely unchanged.
Still others feel the same as above, except they would like to see the system reformed in various ways. How exactly it would be reformed is a conversation that, among slashdotters at least, will result in numerious, boistrous, and often mutually exclusive opinions and suggestions.
Finally, there are some here who ardently support and approve of the patenting system just the way it is.
All of these points of view probably stack up as a minority opinion when taken against all the others.
The myth of a "typical" slashdotter is one being bandied about by Microsoft-paid astroturfers and the like, and has little if anything to do with reality. It is as nonsensical as arguing that a crowd of people who unanimously expect the sun to rise in the east are therefor conformist and can be expected to agree on just about everything else. As with most things, we mostly disagree on the issue of patents and what to do about them and the problems some of us believe they cause. Hell, we even argue about the implimentation of the one thing we all do agree on -- how best to create and nurture free software (BSD vs GPL, Gnome vs. KDE, etc.). If we can't agree on that, it is highly unlikely we'll agree on anything, except maybe that the sun, probably, will rise in the east tommorow, unless of course it doesn't.
The Future of Human Evolution: Autonomy
Why is this patent bad? Google's people invested their hard work and time into developing a page analysis method, and they are more than within their rights to patent it.
Patenting a concept (i.e. the "concept" of banner ads) is relatively questionable business behavior. Patenting an IMPLEMENTATION of that is not - it helps the developers protect their investments, and it forces everyone else to find a better way to do the same thing.
Here Google's authors can rest assured that nobody will steal the fruits of their hard work, and it prompts the rest of the community to come up with a different implementation of a page rankings algorithm that may be better or serve a different purpose.
I'm sure that if someone were to release a competing search engine using Google's algorithm, this entire community would be up in arms about it. So why don't you allow Google the chance to protect their work?
There's nothing wrong with taking out a software patent, in today's crazy world. It's a good way to make yourself look good to investors, for example.
The problem is suing other people based on that patent.
It's really hard to tell without seeing the patent how broad its claims are. On the whole, this patent doesn't seem any worse than a lot of other software patents. Depending on its claims, however, I think there may be some published prior art.
Incidentally, take a look at NorthernLight (www.nlsearch.com); they have a patent on their search folders, again something that is very close to widely used techniques.
On the whole, startups don't have a choice: VCs want patents. Those patents are needed for defense and negotiation with other companies in cross-licensing deals. Almost everybody (other than the lawyers) would be better off if these software patents didn't exist, but as long as the patent office will grant them and courts will enforce them, everybody has to get them.
Software patents are a Bad Thing. It's good to fight to change them. But until they actually are changed, we do have to respect the system.
However, until the patent system is changed, it can yet prove useful. Consider: The GPL uses copyright, which is supposed to prevent people from copying works, to instead allow such copying. Something similar could (albeit more expensively) be done with patents relatively easily. Get the rights to it, then give others those same rights freely.
Last I checked, Google was Open-Source. Since OSS and the typical uses of software patents are antithetical, my guess is that this is what Google is trying to do. It'll apply a GPL-like license to the patent: you may use this technology in your products, provided that either 1) those products do not involve any other patented technology or 2) any other patented technology used is licensed is distributable and usable under the terms of this license.
It's a shame that Google has to do this; patents are expensive. But when you have a system that people twist to knock others out, you have to twist it too in order to survive.
The proliferation of obvious patents (windowing, one-click shopping) is starting to make us lose sight of the core issue: software patents should never have been allowed.
Algorithms cannot be patented. No matter what you may think about that; that's the law. The idea that you can't patent an algorithm, but you can patent the concept of implementing that algorithm in software is simply absurd, and we won't get out of this mess until software patents are rejected.
If the algorithm behind google can't be patented, then it makes no sense to patent the implementation of the algorithm in software. The specific implementation can be protected by copyright law. The *idea* of implementing an algorithm in software is an obvious one.
So even if you think Google has done significant and non-obvious work here, that's irrelevant. The algorithm may be impressive, but that's not patentable. And the implementation may represent significant work, but that's protected by copyright. There's simply no place for patents here.
Now, I'm not really attacking google here. The law is a mess and Google has to play the game. But this madness has to stop soon. The Internet would never have come into being if this mess of patenting every idea in sight existed 25 years ago.
Take the situation of filing a patent in the US. And then you talk about the technology. At that moment the European Patent applications becomes invalid. European Patent applications do not care whether you have filed in the US. The point is that once it is public and no patent has been filed, it become automatically invalid.
"You can't make a race horse of a pig"
"No," said Samuel, "but you can make very fast pig"
Here is the letter I wrote them. I'm not sure what email address you should use for this sort of things; anyone?
Date: Mon, 27 Dec 1999 18:50:02 +0100
From: Vilhelm Sjöberg
To: help@google.com, press@google.com
Subject: Concerning the Patent on PageRank.
X-Mailer: Mutt 0.95.6i
Dear Sir,
I am writing to you, since it has been called to my attention (through the
recent Slashdot article) that you are applying for a patent for the PageRank
measure. I urge you to reconsider this decision. If you feel that that is
impossible, at least consider granting use of this technology to everyone,
without licensing fees.
When I first learnt about Google I was excited, not only about the remarkably
high-quality results it returned, but equally much about the fresh attitude
exhibited in for example the paper "The Anatomy of a Large-Scale Hypertextual
Web Search Engine". Quotes like
"Up until now most search engine development has gone on at companies with
little publication of technical details. This causes search engine
technology to remain largely a black art and to be advertising oriented.
With Google, we have a strong goal to push more development and
understanding into the academic realm."
gave an impression of a company which took true "community responsablity" (to
use a marketing word): rather than being content with some market share, you
aimed to improve the state of the art in web searchers. The academic
background seemed to vouch for a free dissemination of information, following
a tradition much older than IPOs and Market Relations.
With its connections with the academic domain, Google would be clearly aware
of the damages patents on algorithms has done to the level of technology used
in many fields. Research itself might not be hampered (since the academic
study of an invention is not covered by patents), but its adoption in everyday
life is critically hampered.
The clearest example of this is in the field of data compression, where
algorithms like PPM (which has existed for decades) remain unused due to
patent problems. Some more current algorithms _were_ in fact adopted, for
example Ross Williams' variations on Ziv-Lempel coding which were implemented
by GNU, only to be forced to withdraw. Instead, the commercial field remains
dominated by LZ77 (Zip, Gzip), or LZ78 (compress). Imagine if Ziv and Lempel
also had patented their results; then we would still be using per-symbol
huffman coding like the Unix utility pack(3).
The Google founders should be well aware how new inventions in this area must
depend on older; indeed the paper mentioned above itself acknowledges this when
it describes the PageRank:
"Academic citation literature has been applied to the web, largely by
counting citations or backlinks to a given page. This gives some
approximation of a page's importance or quality. PageRank extends this idea
by not counting links from all pages equally, and by normalizing by the
number of links on a page."
Patenting the PageRank would frustrate further development in text searching
and block its actual adoption, in the same way that would have been the case
if the "Academic citation literature" had been covered by patents.
By filing for patent of PageRank, Google has chosen Profit over Progress. If
more users become disillusioned like I have, you might find you will gain
neither.
Sincerely Yours,
Vilhelm Sjöberg
--
-Vilhelm Sjöberg "355/113 -- not the famous number,
vilhelm@home.se but an incredibly good imitation!"
Although software itself is not patentable (and should not be), technologies and algorithms definitely are.
The above-quoted statement is a restatement of a common misconcetion about patent law. It mistates the law, at the same time, by overstating and understating the scope of patentable subject matter. It overstates the scope of patentable subject matter, in that there exists a narrow class of abstract "mathematical algorithms" which are non-patentable. It understates that scope because software, "itself," is most certainly patentable in some cases. Claims directed solely to software have been allowed by the PTO and enforced by the Courts under the United States Patent Act.
The modern standard is basically, "anything under the sun made by man." A patent claim directed to software is patentable, even if it recites a mathematical algorithm, if it constitutes a "practical application of a mathematical algorithm, . . . [by] produc[ing] 'a useful, concrete and tangible result.'" See AT&T Corp. v. Excel Communications, Inc., No. 98-1338 (Fed. Cir. Apr. 14, 1999).
If the lawyer properly framed the claims, it is these days almost frivolous to argue the validity of a software patent claim on subject matter grounds. The only issue is the traditional question as to whether the invention was novel, useful and unobvious, as those terms are used in the Patent Act.
The standard is not whether the invention was obvious to "one 'skilled in the art', e.g. a geek'." To the contrary, the standard is whether "the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains." 35 U.S.C. s. 103(a).
The issue is not whether in hindsight an expert would think the issue is obvious, but whether a dull, ordinary, hack familiar with the art without the benefit of knowing the solution -- or even a requirements specification of the problem -- would find the differences obvious.
So what, then, if we are to charge Amazon with invalidity, is the relevant prior art that makes obvious the one-click solution? And if the solution was so obvious, given the need to do it, why wasn't it done by others before Amazon did it? It is interesting to note that B&N's lawyers couldn't come up with a plausible answer to that question during the preliminary injunction hearings. Where is the "clear and convincing" evidence of invalidity, then, apart from the naked assertion of obviousness so frequently asserted in these letters?
I do not assert that invalidating prior art does not exist. I merely note that no one has come close, so far, to presenting any.
What, pray tell, are patents good for other than restricting your competitors from doing what you have patented? What value does a patent have if there is no threat of a lawsuit? Maybe if you had said "only defensive lawsuits" I would've agreed with you. As it is, I can't distinguish your statement from a hypothetical one like "It's not the sun that's the problem; it's that giant celestial body we orbit around that's the problem."
Even investors, who are overvaluing everything in sight these days, wouldn't touch you on this one.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
I seem to recall a group at Watson Labs working on a similar eigenvector-based ranking system for citation analysis. I believe that research was discussed in a Scientific American article sometime in the last year or so.
Is this the same guy, or am I conflating technologies?
It must not be that obvious, because other search engines have been around for years without using this algorithm.
Now Google uses it, and it actually returns links that make sense because of it. I think it deserves the patent.
--
Win dain a lotica, en vai tu ri silota
There are two successful methods you can choose, depending on your own creativity, to create a search result that will point to microsoft.com.
The first is to go to microsoft.com and find some words that you can string together and take out of context. Remember: Google doesn't preserve the order of your words, and Microsoft.com is one of the most linked websites on the web, so whatever string you put together will likely point to them first before others, and this is even more true if you reference Bill Gates.
To illustrate this, we can put together a simple string of words like Bill Gates eats customers in his office. All of those words are found on Microsoft's site, and lo and behold, microsoft.com is the first site to come up.
The second technique is really just a variation on the first: when Google encounters a word in your search that it doesn't understand, it just drops it and proceeds with the rest of your search. Thus, if you are creative and come up with a word that sounds like an obscene sexual act but which doesn't actually exist in the English language, like for instance "oingoboingo", you can make a lovely search like Bill Gates oingoboingoed your mother. Again, microsoft.com is the first site to come up, after "oingoboingoed" is dropped.
Now you too can explore your own personal brand of juvenile humor.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
Google does not only look at the result pages (micros~1 in this example) for the search terms you entered, but also at the links that point to that page! This is in fact one of the reasons why Google is so good.
I don't think the words evil and satan appear on their homepage, but instead they probably appear on some (or many) links to them.
Look here for more about this.
- Stephan.
Carpe diem.
Damn right, IANAL, but how about this:
A physical device that implements an algorithm is patentable: nobody else can sell such a device without paying royalties.
A list of instructions that carries out an algorithm is non-physical (can be written down on paper, on a floppy, etc.) and is no more patentable than a recipe in a cookbook.
When the list of instructions is written down in an executable file on a PC's hard disk, and that computer then carries out the algorithm, it has for the moment become a patentable device.
HOWEVER, as a practical matter, it is wrong to define a general purpose computer as a specialized device that was designed to carry out one particular algorithm. The PC is really a general-purpose algorithmic simulator that happens to have become very fast.
If I write down a "patented" algorithm on my PC's hard disk, and sell it as a "device", all I'm really selling is a general purpose computer (non-patented) with some instructions on it (non-patented), and BTW it can also run Quake (copyrighted, but non-patented). Do I pay the patent holder 1/2 royalties? If I add Windows (copyrighted, but non-patented) to the "device", I've added thousands of algorithms, so do I pay the patent holder 1/1000th royalties?
I would have to DISABLE all other capabilities from the computer, so that it could only run the original algorithm, and hence become a single purpose "device" that implements it, to be liable under the law.
And as a user, what I run on my PC is my own damn business. I'm not selling it; I'm using it. If I own a web search company and write a google-like ranking engine, I'd be serving up web pages from a general purpose web server. I have bought nothing illegal, and am doing nothing illegal. Get a warrant and all you find is unpatentable components. Would the cop yell "Don't turn the computer off! Without electricity, the evidence will disappear!" ?
As a fantastic example: a crescent wrench is patentable, but a blob of liquid metal that can morph into any shape cannot be called a crescent wrench, if it can also be every other tool in the box. Would the patent holder hire a guy to stand over the blob and wait for it to become a wrench again? "You see? I own that. Pay me!" If the metal blob was in my living room, I'd consider that an invasion of my privacy.
(I've already patented the blob, so you can forget that idea.)
I'm cool like a fool in a swimming p-p-pfft-pool
What I'm wondering is how Google differs from IBM's Clever? According to a Scientific American article, clever determines relevance based on how many pages that are also considered relevant link to a given page. The process iterates several times, re-assigning the weight of each link until the solution "settles out", not unlike a bi-directional associative memory type neural net. That sounds familiar!
At least by my reckoning of the English language. The resulting search puts an American-football team on top.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
Yes, we're all opposed to patenting the obvious (who isn't?) but there are other big problems with software patents that go far beyond obvious stuff.
And the big one is this: software creation doesn't require much capital. It mainly just takes time. In spite of what some people say about "The Software Crisis" or whatever, programmers can be extremely productive. We consider that to be a Good Thing and one of the neat features of this industry. If you have a mind, you can actually create something useful fairly quickly.
Software patents change all that. Every time you write a program or even part of a program, you may be violating dozens of patents and not even know it. Even non-obvious stuff is way too easy to independently invent.
GM can afford to do a patent search on every moving part in their next years car model, but I sure as hell can't afford to do a patent search on every loop in my code. If I have to do that, then the rate that I have to bill would jump from $50 per hour to $5000 per hour. There goes the productivity. That is what makes software patents such a bad thing.
---
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
Patenting an algorhythm is a rediculous concept and needs to be ended as a legal fiction. What would happen if Albert Einstein had patented his theories and formulas? Would every Nuclear powerplant in the world have to pay royalties to his estate?
Actually, his colleague in Chicago, Szilard (sz?) held a bunch of the seminal patents relating to nuclear fission and associated technologies.
And, in case you hadn't heard, every "Nuclear powerplant in the world" did pay many dollars in royalties, directly and indirectly, to patent holders in the underlying technologies during those patent terms.
I have seen dozens of postings in Slashdot recently along the following lines:
Algorithms cannot be patented. No matter what you may think about that; that's the law. The idea that you can't patent an algorithm, but you can patent the concept of implementing that algorithm in software is simply absurd, and we won't get out of this mess until software patents are rejected.
This is a wild overstatement of the law. While it is true that there exists a narrow class of abstract "mathematical algorithms" that are non-patentable per se, the more general statement about algorithms is clearly false. The term, "mathematical algorithm" as used by the courts is not the same as the term is used by "computer scientists," and refers only to claims directed to the most abstract account of a narrow class of mathematical algorithms. This issue has now been well-settled by the courts in Diamond v. Diehr, State Street Bank and the cases, and is no longer considered an open legal question (except, it would seem, on Slashdot).
And even then, the application of an algorithm to a problem (whether in software or otherwise) can be patentable even though the algorithm itself might not be patentable. Indeed, an unobvious selection of a prior art algorithm to solve a particular problem can itself be patentable. The "mathematical algorithm" rule is no more a bar to patents on applications of a "mathematical algorithm" than the "law of nature" rule is a bar to patents on an apparatus that relies on the law of gravity. While it does still have force, the importance and practical import of this rule is still wildly overstated by opponents of software patents.
If you wish to be an effective advocate for your cause, it is imperative that you learn what is, and what is not, the law so that you can criticize it intelligently.
The modern standard is basically, "anything under the sun made by man." A patent claim directed to software is patentable, even if it recites a mathematical algorithm, if it constitutes a "practical application of a mathematical algorithm, . . . [by] produc[ing] 'a useful, concrete and tangible result.'" See AT&T Corp. v. Excel Communications, Inc., No. 98-1338 (Fed. Cir. Apr. 14, 1999).