4 ms·
Ask HN: How will the world change when the PageRank patent expires?
Most Web 2.0 sites including HN would benefit from PageRank or a variant thereof for example.
How do you think the world will change?
- jacquesm 17y agotens if not hundreds of companies are using it right now, but because they're not selling a product based on it there is no infringement. You have to prove that, as long as you can't take apart some machine to prove infringement it is just another fishing expedition. And even if you did, I'm not sure if that would constitute infringement (since you are not selling an apparatus containing the infringing item). Maybe one of the legal eagles here can fill that one in ?
- fatdog789 17y agoThe use of the patented device/method/whatever, whether or not it is sold, is the infringement. Unwitting infringers are not subject to very harsh penalties, if any (unlike ignorance of the law, ignorance of a patent is a mitigating factor, b/c people are not required to research the patent archives for everything they do or make). Knowing infringers can be hit with the hammer of holy justice. Example: patents on custom protein structures. Note that pure research, i.e., at universities, is usually considered a defense b/c it is in the public interest to encourage actual research, but it is a defense that must be proven in court. However, research with any sort of commercial motive will usually not avoid infringement liability. The reason that the "hundreds" of companies can get away with it is b/c it isn't worth the cost to the patent holder to pursue them in court -- damages would be minimal, and the chance that a defendant could find something that would overturn the patent (i.e., prior art) is too high a risk if there are at least some companies willing to pay for a license.
- wmf 17y agoIANAL, but using a patent is grounds for infringement. Unlike the GPL, patent law has no SaaS loophole.
- evgen 17y agoTrue, but how would you prove it? A service just gives you the results, it does not tell you how it comes up with the results. Someone looking for infringement would need access to a similar base dataset and would have to run a lot of queries to develop enough proof to get any potential lawsuit off the ground. The only path I could see for such a lawsuit to develop would be one where a disgruntled employee provides enough evidence to get things rolling.
- alain94040 17y ago1. You sue. 2. You go through discovery, comb every line of code of the person you are accusing, until you find what you are looking for It's really that simple. All it costs is money (for both sides unfortunately).
- jacquesm 17y agono, it's not that simple, that is a fishing expedition. No judge would allow that during discovery.
- fatdog789 17y agoQuite simple: a plaintiff is not required to have complete proof of the defendant's wrongful acts prior to suing them in court. The burden would be too high, and most aggrieved plaintiffs would be unable to recover the harm caused them by the defendant. Plaintiffs have the right to get "discovery" to prove their claims because the defendant is usually the only person/company with evidence of the defendant's wrongful acts. A fishing expedition is where the plaintiff cannot articulate any particular type of evidence sought, or how such evidence would be useful/necessary to its case, or how such evidence would result in evidence it could actually use. Example: P sues D for patent infringement, demands in discovery D's head researcher's notes. Not fishing. Example 2: Same, except that P demands the CEO's call logs. Fishing, b/c not reasonably related to the patent infringement claim. Example 3: Same as Ex 2, except that P alleges that the CEO knew about the patent b/c he spoke with a former employee of P before launching D's infringing product. NOT fishing, b/c now it's shown to be relevant. How to tell the difference: If you can't articulate something relevant to your claims that the discovery could reveal, it's a fishing expedition.
- vaksel 17y agomost likely it won't change a bit
- rs 17y agoAgreed. Having PageRank in your product is one thing, but using it well and scaling it up (like Google has done) is a way different (and probably much more difficult) matter.
- neilk 17y agoGoogle itself has largely abandoned PageRank as described in the original paper. It's only one of the signals used to rank web pages. I don't understand your assertion that Web 2.0 sites would benefit from PageRank. And in any case, the patent is not what was holding them back.
- qeorge 17y agoSo they say. It may be "only one of the signals", but its the biggest by a long shot. Don't believe what Google tells you about its algo, test it yourself. Remember, its in their best interests for everyone to think their algorithm is so complex that it can't be spammed or replicated. Neither is true.
- onreact-com 17y agoNo, by now domain age and authority are most probably the single most important ranking factors of Google, among 200+ other so called "signals". I can see it in my SEO work. No domain age and authority means you can't rank in results even with substantial PageRank.
- vannevar 17y agoNot much, PageRank was never the brilliant innovation it was made out to be. It was a nice improvement ca 1998 but Google has spent most of the last ten years throwing patches at the problem of people gaming the system. They've been the Red Queen, running faster and faster to stay in place. In terms of actually finding what you're looking for, I don't think Google today is much better than it was then, which is maybe twice as good as Alta Vista 1998. Search engine scalability has come a long way, but search engine results have improved little.
- Anon84 17y agohuh... there will be one more expired patent? Even at this point (30 or 40 years) from that day, PR is, practically, only of academic interest.
- throw_away 17y agopatents now expire in 20 years, so http://www.google.com/patents?vid=6285999 http://www.google.com/patents?vid=6285999 expires in 2018. agreed, however, that the world will be so vastly different by then that this won't be as significant as, say, the rsa or lzw expirations (unless google's lawyers get their teeth into anything).
- qeorge 17y agoEvery major search engine (Google, Yahoo, Bing, ~Ask) uses a form of "PageRank" for their rankings. That is to say, all major search engines use the a graph as their primary ranking metric (other than relevancy, obviously). I don't know how the patent plays in, but its fair to say "the cat is out of the bag" that a weighted link graph is (currently) the best way to algorithmically determine a given website's authority. Disclaimer: I don't work for a search engine, but I am an SEO.
- ilkhd2 17y agoVery lightly. more small vendors will start selling inmplementations...
- brk 17y agoPageRank is only interesting to people who do not understand it.
- amichail 17y agoCould you elaborate? Do you disagree that using it in social news sites would result in an improvement?
- brk 17y agoYes, I disagree, but let's approach this from a different angle. What do YOU think is so unique and valuable about pagerank and its applicability to social news sites? From where I sit, it's just an algorithm to spit out a relative 1-10 ranking of a given pages authority.
- amichail 17y agoYou can use it to compute the importance of submissions and users in a way that is less susceptible to manipulation. Submissions by more important users would themselves be more important before anyone else votes on them.
- andrewljohnson 17y agoI don't think PageRank is what set Google apart. It was multi-word tokens.
- andrewljohnson 17y agoAccording to Matt Cutts, PageRank is the main way Google orders pages for crawling. If you have high page rank, you get crawled first. mat didn't say this, but beyond that, I think the reaction of users is by far the biggest determinant in which pages get top ranked.
- onreact-com 17y agoThat's true but you can make Google crawl your site even without PageRank. Think XML sitemaps and blog pings.