5 ms·

Short and very informative, though I wondered why patents are no concern for a competitor of Google (they have various broad and trivial patents on using histor

Short and very informative, though I wondered why patents are no concern for a competitor of Google (they have various broad and trivial patents on using historical data, user feedback, moderation for better search results). As the author wrote, aiming at feature parity with Google is probably a bad idea. DDG seems to do this right: they have some better features and one distinct advantage over Google with their stance on privacy. Perhaps Google will reach a certain creepiness factor with all the collected data they're factoring into search results (e.g. I have this suspicion that content from gmail is used too) and at some point it might work against them even for the average user (or am I just biased?). One "feature" that Google has failed to do properly is custom "site search", i.e. providing a search interface for arbitrary websites. Perhaps this is another opportunity for a new competitor or even Bing (though their "site search" was shut down in 2011 apparently), there is plenty of demand (and lots of crappy custom solutions).

antics 13y ago
Just a small note about sitesearch, Google are definitely not the market leader there. They don't seem to care much about this feature.
- wslh 13y ago
  I wrote an opinionated blog post exactly about that saying[1]: "When each site understands its own data better than Google, its internal search results will surpass Google’s. Google will no doubt continue to provide better global results, but the two-tiered search would decentralize efforts to improve algorithms." I see it as one of the Achilles' heel of Google. Even if what they are doing seem impossible to copy. - [1] Letters from the Future: Challenging Google’s Search Engine: http://blog.databigbang.com/letters-from-the-future-challenging-googles-search-engine/ http://blog.databigbang.com/letters-from-the-future-challeng...
  - lazyjones 13y ago
    Currently, it is very easy for websites to provide better search results (but still not good) compared to Google and other search engines like DDG due to the limited crawling, except perhaps for the top few 1000 websites with small enough number of pages/URLs so that at least Google can keep them fully indexed. This is also where dedicated "site search" providers can beat these search engines easily: they can offer a full crawl and depending on the size of the resulting index, they can bill the website owner (e.g. by storage used in GB or number of URLs).
    - wslh 13y ago
      I don't think so. Google search gives more accurate results than StackOverflow itself.
- PaulHoule 13y ago
  Site search is an entirely different problem than searching the web, in fact, really good site search for 100-1000, 10,000-100,000 and 1,000,000+ are entirely different problems.
salient 13y ago
Because Google usually doesn't use its patents aggressively. They will use them defensively, though, or even as a counter-attack. That's why I think it's incredibly stupid of Microsoft to be attacking Google with search patents through Rockstar. If they piss off Google enough to start using their search engine patent trove against Bing, it may be game over for Bing. Microsoft should already consider itself very lucky that Google isn't as spiteful and vengeful as Apple, or Google would've used them since day one of them going after Android OEMs. That being said, I think the only solution to beating Google is developing something disruptive, and I mean that in a very classical and traditional sense (not the way disruptive is used to mean anything these days). It means you have to build a type of a search engine (or search solution) that Google not only has no interest in pursuing (due to conflicts of interest/opposite business incentives), but probably can't pursue, or not without employing dramatic shifts in how they do search (which again, makes it even less likely for Google to pursue and try to beat this new "disruptor"). And no, don't look at Bing to provide that. They're a "direct competitor" to Google, and they will always be, because they want to be "in the exact same business" Google is, and therefore, I don't think they'll ever radically change the way they build a search engine compared to Google. So what can such a disruptive solution be? I can't say for sure, because it probably needs more factors/benefits on its side than just one, against Google. Social "may" be a way to disrupt Google, although I think Google is already rapidly adapting to this attack vector. I think another could be "extreme-privacy", which Google wouldn't pursue anytime soon without really rethinking how its search business works. Disruption also means attacking incumbent companies with much fewer resources through innovative business models and cost structures. Do something that would make Google's billions of investment in infrastructure irrelevant. A P2P search engine like YaCy or Faroo seems to be part of that solution. Perhaps what can/will kill Google is a crowdsourced super-search engine, that ends up surpassing even Google in most capabilities, because Google would be just one company, fighting against the whole resources of the world. As in classic disruption, you also have to consider this new type of search engine will be much poorer than Google in some competition factors (at least initially). I think that is the "relevance" factor. As the op mentions, it's going to be very hard for anyone or anything to beat Google in relevance. They have too much of a headstart. But perhaps there will be such a thing as "good enough" in search, too. Maybe we're not there yet, but perhaps we will be in 5 years, or 10 years. Until a few years ago we kept wanting faster and faster Intel processors. Now we think mobile ARM processors are more than good enough for most of our daily activities. A crowdsourced P2P search engine powered by billions of people could perhaps reach that point, too, while offering some of that "extreme privacy" and other benefits that Google can't offer. Still, how do you get that first "10 percent" of users, that end up changing the paradigm, to start using a service that's much less relevant than Google today? You do it in the same way other disruptive technologies win out against incumbents. You promote the other benefits, that Google can't and won't match - benefits that will be important to those "first 10 percent", until you reach critical mass, and then you should be fast approaching that "good enough" relevance, too, and get even more users after that.
w1ntermute 13y ago
> I wondered why patents are no concern for a competitor of Google (they have various broad and trivial patents on using historical data, user feedback, moderation for better search results). How exactly are server-side patents enforced? When the hardware on which the infringement is occurring isn't exposed to the public, how can anyone verify its presence? I know that for some (many?) of Google's search patents, there might be an indication on the client side if infringement were occurring, but I imagine that there are also many patents for which this isn't the case.
- throwawaykf03 13y ago
  You don't need convincing proof, just reasonable suspicion to file a lawsuit. (There's something called Rule 11 to discourage frivolous lawsuits, but it's a pretty low bar.) You can then determine actual infringement during discovery. This has some problems however: 1) As you can imagine discovery may not turn up any actual infringement, making it all a huge, expensive waste of time. 2) The unfortunate reality of patent lawsuits is that if you can't prove infringement just by looking at something, you might as well have already lost. Anything that requires expert witnesses to provide input on often comes down to which sides' witness the jury finds more likable, and that's pretty much a roll of the dice.
  - w1ntermute 13y ago
    Couldn't "discovery" be abused to get a peek inside other companies' systems?

antics 13y ago

Just a small note about sitesearch, Google are definitely not the market leader there. They don't seem to care much about this feature.

wslh 13y ago
I wrote an opinionated blog post exactly about that saying[1]: "When each site understands its own data better than Google, its internal search results will surpass Google’s. Google will no doubt continue to provide better global results, but the two-tiered search would decentralize efforts to improve algorithms." I see it as one of the Achilles' heel of Google. Even if what they are doing seem impossible to copy. - [1] Letters from the Future: Challenging Google’s Search Engine: http://blog.databigbang.com/letters-from-the-future-challenging-googles-search-engine/ http://blog.databigbang.com/letters-from-the-future-challeng...
- lazyjones 13y ago
  Currently, it is very easy for websites to provide better search results (but still not good) compared to Google and other search engines like DDG due to the limited crawling, except perhaps for the top few 1000 websites with small enough number of pages/URLs so that at least Google can keep them fully indexed. This is also where dedicated "site search" providers can beat these search engines easily: they can offer a full crawl and depending on the size of the resulting index, they can bill the website owner (e.g. by storage used in GB or number of URLs).
  - wslh 13y ago
    I don't think so. Google search gives more accurate results than StackOverflow itself.
PaulHoule 13y ago
Site search is an entirely different problem than searching the web, in fact, really good site search for 100-1000, 10,000-100,000 and 1,000,000+ are entirely different problems.

salient 13y ago

Because Google usually doesn't use its patents aggressively. They will use them defensively, though, or even as a counter-attack. That's why I think it's incredibly stupid of Microsoft to be attacking Google with search patents through Rockstar. If they piss off Google enough to start using their search engine patent trove against Bing, it may be game over for Bing. Microsoft should already consider itself very lucky that Google isn't as spiteful and vengeful as Apple, or Google would've used them since day one of them going after Android OEMs. That being said, I think the only solution to beating Google is developing something disruptive, and I mean that in a very classical and traditional sense (not the way disruptive is used to mean anything these days). It means you have to build a type of a search engine (or search solution) that Google not only has no interest in pursuing (due to conflicts of interest/opposite business incentives), but probably can't pursue, or not without employing dramatic shifts in how they do search (which again, makes it even less likely for Google to pursue and try to beat this new "disruptor"). And no, don't look at Bing to provide that. They're a "direct competitor" to Google, and they will always be, because they want to be "in the exact same business" Google is, and therefore, I don't think they'll ever radically change the way they build a search engine compared to Google. So what can such a disruptive solution be? I can't say for sure, because it probably needs more factors/benefits on its side than just one, against Google. Social "may" be a way to disrupt Google, although I think Google is already rapidly adapting to this attack vector. I think another could be "extreme-privacy", which Google wouldn't pursue anytime soon without really rethinking how its search business works. Disruption also means attacking incumbent companies with much fewer resources through innovative business models and cost structures. Do something that would make Google's billions of investment in infrastructure irrelevant. A P2P search engine like YaCy or Faroo seems to be part of that solution. Perhaps what can/will kill Google is a crowdsourced super-search engine, that ends up surpassing even Google in most capabilities, because Google would be just one company, fighting against the whole resources of the world. As in classic disruption, you also have to consider this new type of search engine will be much poorer than Google in some competition factors (at least initially). I think that is the "relevance" factor. As the op mentions, it's going to be very hard for anyone or anything to beat Google in relevance. They have too much of a headstart. But perhaps there will be such a thing as "good enough" in search, too. Maybe we're not there yet, but perhaps we will be in 5 years, or 10 years. Until a few years ago we kept wanting faster and faster Intel processors. Now we think mobile ARM processors are more than good enough for most of our daily activities. A crowdsourced P2P search engine powered by billions of people could perhaps reach that point, too, while offering some of that "extreme privacy" and other benefits that Google can't offer. Still, how do you get that first "10 percent" of users, that end up changing the paradigm, to start using a service that's much less relevant than Google today? You do it in the same way other disruptive technologies win out against incumbents. You promote the other benefits, that Google can't and won't match - benefits that will be important to those "first 10 percent", until you reach critical mass, and then you should be fast approaching that "good enough" relevance, too, and get even more users after that.

w1ntermute 13y ago

> I wondered why patents are no concern for a competitor of Google (they have various broad and trivial patents on using historical data, user feedback, moderation for better search results). How exactly are server-side patents enforced? When the hardware on which the infringement is occurring isn't exposed to the public, how can anyone verify its presence? I know that for some (many?) of Google's search patents, there might be an indication on the client side if infringement were occurring, but I imagine that there are also many patents for which this isn't the case.

throwawaykf03 13y ago
You don't need convincing proof, just reasonable suspicion to file a lawsuit. (There's something called Rule 11 to discourage frivolous lawsuits, but it's a pretty low bar.) You can then determine actual infringement during discovery. This has some problems however: 1) As you can imagine discovery may not turn up any actual infringement, making it all a huge, expensive waste of time. 2) The unfortunate reality of patent lawsuits is that if you can't prove infringement just by looking at something, you might as well have already lost. Anything that requires expert witnesses to provide input on often comes down to which sides' witness the jury finds more likable, and that's pretty much a roll of the dice.
- w1ntermute 13y ago
  Couldn't "discovery" be abused to get a peek inside other companies' systems?