Dear Google: You keep using that word...

Google's Jonathan Rosenberg wrote a paean to 'open,' in which his company's commitment to 'open' is pitched at great length. The most remarkable paragraph, however, is the one dealing with things that Google keeps closed:
While we are committed to opening the code for our developer tools, not all Google products are open source. Our goal is to keep the Internet open, which promotes choice and competition and keeps users and developers from getting locked in. In many cases, most notably our search and ads products, opening up the code would not contribute to these goals and would actually hurt users.
How odd that of all the products Google would be forced to keep proprietary by its commitment to an open internet, it just happens to be the ones that make it all of its money.


    1. they also seems quite a long way from ‘don’t be bad’.

      has anyone here read their books settlement?

  1. I’m A-OK with them keeping their search algorithms a secret. The humungous revenue they generate from profitable endeavors has helped to bring me gMail, google maps, google voice, and the ability to use google as a verb.

    Is this post a statement of fact or a condemnation?

    Sure open source is great. All of my primary programs (except my OS) are open source: FireFox, ThunderBird, Filezilla, VLC, Open Office. But just because I love getting open source software, that works excellently, for free – doesn’t mean I think all code should be public/free.

  2. It’s not a matter of evil and good.

    Don’t get me wrong, I’m almost always a proponent of open-source. But I have to admit that it makes sense that they want to keep their revenue generators close to the vest.

    If they opened up Google Search, then people would immediately know how to game the system to get their pages at the forefront of the search lists. And if they opened up Google Ads, people would immediately know how to game the system to illegitimately pay less or get paid more.

    Now, it could be said that if they built their products well enough, those events could be avoided, certainly. But I don’t blame them for taking extra care.

  3. It’s interesting, with all the recent talk at google about how unimportant anonymity is, that their “open”-ness is quite selective. What a luxury.

  4. With the amount of energy and money that goes into gaming the google rankings I think they’re 100% right — it’s better for the user if the exact algorithms aren’t floating around to game.

  5. Honestly, would any company in any industry be expected to share its trade secrets? I feel sorry for those who think the big internet companies are part of vast conspiracy to control us with personal information. True privacy and anonymity were all but gone 20 years ago, anyway (started by the credit card industry, NOT the internet). Even the oldest of our living citizens most of whom have no internet access) can be found referenced somewhere online. I love Open-Source and Fair Use Laws, but I also believe a company has a right to protect its profits. You also failed mention that Google has pledged $20M in charitable donations this Christmas, to be distributed among some 2 dozen organizations, world-wide. And no, I am NOT a Google employee.

  6. (Edit: I just noticed eain and unicorn breath’s comments – this is sort of in reply to them)

    To be fair, the following two sentences do somewhat justify this statement:

    >>The search and advertising markets are already highly competitive with very low switching costs, so users and advertisers already have plenty of choice and are not locked in. Not to mention the fact that opening up these systems would allow people to “game” our algorithms to manipulate search and ads quality rankings, reducing our quality for everyone.< < On the surface this reasoning seems to make sense, even if the person making the argument has outside motivations, but I've been reading alot of Schneier lately ( The first thought that springs to mind is whether or not this is the locksmith’s code of silence all over again? Namely that if Google keeps its algorithm secret your average user has the presumption of fairness and is confidant that noone is juking it – while in reality there are most certainly people that have figured it out, and to them the system is no more difficult than a two-pin tumbler, something I’ve picked with home-made aluminum tools.

    If they open-sourced this code then hackers and mathematicians would have a chance to analyze and point out flaws; website owners would be able to detect whether they were being given fair-play or not; and any black hats out there would find successive patches ruining their fun, patches developed by online communities, not just the limited staff at Google.

    As an analogy, I like to think of it as D&D – the best way to prevent cheating is if all of the players know all of the rules for everybody else.

    The real argument behind all of this, is that since (presumably) these algorythms can’t be patented, for Google to release them would give their competitors an unfair advantage. They’ve busted their asses to make a good search engine, and I don’t really mind if they’re the only ones to profit from it. Same as the Drug companies, as far as I’m concerned – all the power to them.

    So all of this is fine… but I really don’t like the mendaciousness to all of it. Aloisius is right: “Do no Evil” is not the same motto as “Do Good”.

    1. In all honesty, everyone knows pretty much everyone else’s algorithms. It’s not the algorithms so much. Most of the algorithms is some sort of machine learning algorithm. Knowing the algorithm isn’t anything. Most of them are just weighted sums. Knowing the weights, and what exactly is being summed up is the secret part, and that’s fine. It doesn’t hurt the science, and any way everyone has a pretty good idea on what Google is using.

      Don’t believe me? Take the Search Engine Taste Test. Everyone’s results are pretty much the same.

    2. @10

      Aurini (and many others), you are conflating Google’s search algorithm with a piece of software. Software tends to benefit from openness because, yes, it allows more people to find and fix bugs and holes. Knowing the software’s source code does not necessarily allow you to “game” the program to the detriment of others. A search algorithm is not something that always benefits from an open source community providing feedback – there may or may not be flaws in Google’s algorithm, but if they let everyone know all the details of it, everyone would be able to use sneaky tricks to game the system to get ahead.

  7. In many cases, most notably our search and ads products, opening up the code would not contribute to these goals and would actually hurt users.

    users, a.k.a. the unwashed masses.

  8. Do we need to be reminded that the whole “don’t be evil” thing is just a guideline, and is interpreted to mean “don’t be evil for no good reason”? They’re quite happy to accept “lesser of two evils” (see their policy on China).

    1. They’re quite happy to accept “lesser of two evils” (see their policy on China).

      You. F-ing. Moron.

      If a company wants to operate anywhere, they have to play by local rules and abide by local laws. China has very strict internet regulations. Google has to abide by them if they want to do business in China. Considering the internet is booming in China, I would say its a pretty good business move for a giant like Google to set up there.

      If you have a problem with how the internet is regulated in China (as I do) then you should direct your anger at the people responsible (Chinese Government), not the companies whose hands are tied.

      If you are refering to the fact that does not index banned material as proof that they are evil, then you really need to do your research. The fact they do this is actually a benefit to the world, as it gives us and Chinese web users a means of checking what content is and isnt banned by the Chinese government. People can still visit in China. By typing the same search term into both & and then comparing the differing search results we (and Chinese web users) can get a picture on what type of content is blocked. That information is better than no information, no matter how you look at it.

      If you want to direct you anger at some US company, then how about you try Cisco Systems and Juniper Networks — the 2 US companies to help carry out its network upgrade in 2004.

      Or how about Yahoo or Skype? Who actively assisted the Chinese Government to spy on Skype & Yahoo conversations.

      Misdirected anger from a person who has nfi.
      You’re really not think-tank material, are you?

  9. I think the people who are most upset about Google’s secrets are the cyber-libertarians who think all information (in order of importantance; music, movies and t.v. shows) should be free.

  10. Dear Rob: Inconceivable that you’re wrong.

    Had Google claimed its ad algorithms were “open” but refused to divulge their innards, then yes, I would agree they are misusing the word. But clearly they are not.

    Anybody want a peanut?

  11. This is no flame against Rob, but I actually disagree with the entire thrust of the post. Open software allows for a community to develop the application so that others can use the improved application as they wish. Search and Adserving, on the other hand, is a service, not software, which needs to be optimized to run on Google’s hardware and internet backbone, to seamlessly interact with Google’s other data-mining algorithms, and to work in conjunction with Google’s account system.

    I don’t see how anyone can act insulted that Google isn’t opening up the software that power the services they provide. Not only are those two areas of software development strictly useless unless you have the petabytes and petaflops of information in addition to the market reach, but open-sourcing that software would make it easier for players to game the system. Normally I’m against security through obscurity, but there just doesn’t seem to be much justification for open source here.

  12. Search and crypto are not the same thing. With cryptography, there’s a strong argument for open algorithms, and the open algorithms have proven highly resistant to attack, while attempts to produce proprietary algorithms have usually failed. But no one knows how to write a search algorithm that will return good results if everyone (including spammers and con artists) knows the algorithm. Given this fundamental difference, Google and its competitors are going to need to keep the details secret, for now.

  13. “How odd that of all the products Google would be forced to keep proprietary by its commitment to an open internet, it just happens to be the ones that make it all of its money”

    That’s because the good people at Google are a bunch of hypocrits.

  14. I don’t know. It doesn’t really bother me much. They are a For-Profit business, so they keep their revenue generating code closed. I… I really see no particular problem with that. A popular restaurant is unlikely to hand out flyers with the recipe for their most popular dish. Google is not a non-profit, so they have no responsibility to do things that will hurt their profit margin. The fact that they choose to use the profits from the few closed services they offer in order to provide a significant number of open services seems pretty close to wonderful to me. And they aren’t even lying about what they keep secret! So why the disapproving tone? It doesn’t seem like they’ve done anything wrong here.

  15. Google always has some innovative thinking. This is the most popular site there is no doubt about it. Sometimes it takes decision which pretends to be foe to the visitors and for the members but there intention is to serve society without any hazard. I believe that Google will emerge with a bang.

  16. I see a lot of people missing the point.

    Open isn’t ‘open to a point’. Open is open.

    “Mostly open” is still “occasuionally closed” – as illustrated by the ironically named Store 24 chain, many of which are open 18 hours a day.

    1. No, I disagree. “Open” isn’t like “pregnant” or “dead.” I see no reason why a very large organization that does a great number of things can’t be “partially open.”

      Your point really breaks down when you look at it the other way. If an organization is either Open or Not Open, then it is also either Closed or Not Closed. Well, many, probably most, of the things Google offers are open, so it seems clear to me that they can’t be completely closed.

      You seem to view it as a hard edged dichotomy between a utopian, free-code collective where every piece of software is open to everyone, and a villainous, clenched-fist monolith of corporate tyranny. I don’t think those are our only options, though.

      1. etho, I think your point really breaks down when you turn the argument away from what words google uses, and towards what words I should use.

        And yes, they should use ‘closed’ rather than ‘open’. To make my point really really clear -> “Mostly not a rapist” is still “a rapist”. Or if you prefer a less obnoxious approach “mostly bug free” is still “buggy”.

        You seem to view it as a hard edged dichotomy between a utopian, free-code collective where every piece of software is open to everyone, and a villainous, clenched-fist monolith of corporate tyranny. I don’t think those are our only options, though.

        Where do I sem to believe that? I’m just being a fascist about people using sunny words to hide their motives. I’m not saying they’re evil… I’m just saying they’re patting themslves on the backs a bit too hard while missing the mark they’re bragging about hitting.

        Not unlike how I feel about people using “free country” to describe a nation who deprives people of due process.

        It IS semantics, but those are important, ESPECIALLY to a computer company. GIGO, after all.

  17. This seems like a cheap shot, Rob.

    Maybe you’ve got some theory about how they could do things better, being more open in a way that yields some public benefit and isn’t obvious suicide for them. In which case, speak up.

    But as a long-time supporter of open approaches, I feel like Google does pretty well for a tech behemoth. There are an ocean of dubious things they could do and don’t. They have open-sourced quite a bit, including Chrome and Android, as well as smaller things like Guice. And they are the largest source of revenue for the Mozilla Foundation.

    Overall, I’d say they’ve earned the right to toot their own horn a bit on this.

  18. Nobody is denying Google the right to keep their core secret but by the same token one can hardly then make an argument for other folks to embrace openness by giving up their core differentiators. That’s a self serving argument that only benefits Google business model while dressing them up as some sort of champions of openness.

  19. I am curious where all these people are who so stringently insist that Google must release all their stuff. I see many people arguing against them, but I don’t see anybody taking that position.

    I suspect people are just airing their prejudices without regard to the subject at hand.

  20. Anybody here actually buys that anyone would aggregate that much data on that many people to Not Be Evil?
    I have a server park to sell you in Cali. Real cheap.

  21. I would say Google is among the top 10 contributors to open source software, which is good.

    By giving away their business secrets, they would no longer be able to generate the income that they do, and would be no longer able to contribute to open source in the same manner.

    I see nothing wrong with a company keeping its business secrets as that. That they’re contributing anything to open source whatsoever is commendable. Most companies don’t.

    I also don’t think this has much to do with being good or evil – how they use people’s data is another issue altogether.

  22. Dear Rob: I do not think it means what you think it means.

    Believing that openness is good in general doesn’t equate to to believing it’s good in all specific cases.

  23. Google is going in the right direction opening up their solutions. Android, Chrome and several more projects are extremely valuable to their user base. Of course, they could always be doing a bit more, say opening up (parts of) Gmail, Google Apps, GFS or even their web serving stack, but ads and search don’t really need to be more open.
    The search part is debatable. Arguments in favour of opening it up are transparency (meh), using it as a learning resource and actually using the code in new solutions. Transparency is also the biggest contra: as long as SEO is around, opening up would make the system ripe for abuse by “optimizers”, and this time they’d actually know where to efficiently attack the index. Learning from the actual algorithm strikes me as rather unrealistic; in an academic setting, a more basic implementation of PageRank would probably teach the principles in a more graspable fashion. For some enthusiast tinkering, it also strikes me as too big, existing open source search engines ought to be better suited. In my opinion, as long as it’s not absolutely bulletproof, the costs outweigh the benefits.

    In ads, I don’t really see the use-case in opening up. The system is built to make money by serving up ads. It’s use is limited to, well, companies that make money by serving up ads. Coincidentally, this is the way Google makes the vast majority of it’s money. They probably invested tens of man-years into that platform, why suddenly give it away to what only can be competition? This also benefits the web as a whole by keeping the number of I’ve-downloaded-Google-AdSense-and-am-serving-ads-for-cheap-now startups. Imagine the turnaround time for AdBlock lists if all it took to set up an ad serving service were two hours.

  24. Sigh. It seems most here are missing the point. No one is calling fir Google to open source their important code. But the idea is that it’s hypocritical to have an exec write an essay extolling the need for everyone else to do so. You guys need to make all of your products open because it’s for the greater good, but we don’t because it might hurt our profits.

  25. I’m with google on this. Opening search algorithms would only make a mess, everyone would know how to get their page on top, it would no longer be about the quality of the page, but rather the abuse of the google algorithms.

  26. I’m usually one of the first to jump on the corporation-bashing bandwagon (such a fun wagon!) but I don’t have a big problem with Google using its own property to turn a profit (and stay in business).

  27. The make a profit being “open” about your internet history with the police and government and anybody else who pays them for it.

    But they would rather not talk “open”ly about that….would not want to aid the “criminals”, by being “open” with them as to who’s watching, why or how, eh?

  28. The problem isn’t that google keeps its money-making stuff proprietary, which is A-OK by me–it’s that it’s being deceptive about *why* it keeps that stuff proprietary.

    We are so used to corporate spin of these things that it’s becoming invisible. You can’t point out the spin (here, that closed is necessary for the health of the open internet) without people thinking you’re also criticizing the thing being spun (closed is necessary for the health of google’s business).

  29. Search algorithms aren’t the only thing Google keeps secret.

    Last time I checked, Google search doesn’t search web page source code. (There’s a “Search Code” feature, but it searches other kinds of code, not web page source code.)

    This bugged me at first — I wanted to find pages made in Dreamweaver, by searching for fragments of Dreamweaver-specific code — but I’ve come to assume that Google doesn’t want us to search web page source code.

    If we could search the source code, what’s to stop us from writing our own ranking algorithms, bypassing Google’s rankings altogether?

    1. “Last time I checked, Google search doesn’t search web page source code. ”

      They used to at one time , but putting search terms into the code became a way to spam the results so Google stopped letting you search for things like javascript routines.

      It’s a drag if you are looking to find code, but not so much if you are looking for documentation.

  30. I’m cool with them keeping the search and ads products. One they need the adds to stay alive. and keep people employed.
    Two if they has the code open source companies like MS would walk right in and steal and crush the company. For a company Google is very Open and I love that this company is able to stay alive in a cutthroat environment. I use Google for everything from Work, Searching for a recipe for dinner and how to get to a location I have never been to. This is all free to the user.

  31. If Google opens the search algorithm people will learn how to game it and the results would just turn into spam.

    With the adsense/adwords programs – let’s face it Google is a business and they want to make money. The are a public company and have to keep profits up for shareholders.

    Keeping the web open has nothing to do with sharing proprietary code.

    Don’t be evil just means don’t screw over the public.

  32. If they opened the search code, there might be temporary pain from people gaming the system, but that would be buttoned up really quickly and in the long term the search experience would be far, far better.

    As Rob says, this is not about protecting users, it’s about protecting revenue. Which is fine, but just tell the damn truth. It’s not that complicated or difficult.

    1. As Rob says, this is not about protecting users, it’s about protecting revenue. Which is fine, but just tell the damn truth. It’s not that complicated or difficult.


      But in a world where the public’s spending for a debased holy day makes or breaks the entire economy a lie has to be a REALLY BIG LIE for it to be called a lie, and honesty has to be coated in sugar for it to be anything other than an attack.

  33. “How odd that of all the products Google would be forced to keep proprietary by its commitment to an open internet, it just happens to be the ones that make it all of its money.”

    Circular reasoning. It’s possible they’d make much more money by keeping some of their open products closed. Admittedly, none comes to mind, but 10 years ago, who thought there was money in search?

  34. Google isn’t in the business of telling the truth, any more than they’re in the business of benefiting the internet or humanity in general. They’re in the business of business: getting the paper by any means necessary. Principles and slogans are just convenient decorations.

    That’s not to say the heads of Google are bad people – this is just how capitalism works. It reduces everyones ideals and aspirations to functions of a market equation, and usually forces us to drop them in favor of better performance. There’s no reason to believe Google would be any different, and there’s no reason to believe we would do any different in their position.

    Doesn’t mean they should get a free pass, but as the dozens of unimpressed commenters demonstrate, the problem is deeply systemic.

  35. Search and ads isn’t just how Google makes money, it’s how most websites make money. Making them as hard as possible to exploit keeps the playing field level, as well as making them more useful to the people using them, by “promoting choice and competition.” Makes sense to me.

    Also, being a fan of all the free, no-obligation, open stuff Google gives me, I might say their secure revenue stream is itself indirectly but greatly beneficial to an open internet. I mean, until they officially become evil.

  36. It’s an interesting argument you make. I find one central problem with it:

    In D&D, everyone is playing to win largely the same prize. Get the treasure, slay the dragon, etc.

    In the Internet, people have disparate goals. People develop web sites for all sorts of reasons. Some people are playing the game, some aren’t. Most top hits on google aren’t due to SEO, but rather just good content and popularity.

    However, if the search code were made public, the tilt of what Google pointed at would be shifted massively towards those people who took the time to study and understand the algorithms, and away from the organically good ranking content.

    In a good example of game theory, web sites in competitive fields would have to devote more and more of their time to this SEO with diminishing results.

    The end result would be less organic relevance of top results and people spending ever more of their finite time and resources trying to game the system instead of creating valuable content.

    I’m all for open source, having recently zapped the windows partition on my only PC. And I’m not exactly Google’s #1 fan. But, google has both a right and a very good justification to keep some of these core technologies secret.

  37. So a search algorithm is suddenly more culturally important that the books and poems of a nation or the employment that other software providers provide?

    Google only wants open if it doesn’t affect their financial health. Meanwhile they continue market dumping on software providers for email, mapping, accounting, word processing and now time management.

    Add to this stealing book publishers content so they can jack up revenue streams. They need to learn to coexist and not just do what suits them to flick on a few more adverts.

  38. Google’s domination in the search market (although based on the excellence of its product) is an issue. I would love to see Internet search decentralized, and open search software would lead to all sort of innovations. I wonder, however, how easy or how hard it would be for someone to start a search engine from their garage using open code. Would it require prohibitive processing power, or is it something someone could do with a fair number of used, modded Xbox 360s? If the barrier of entry in the search arena is too high, we have a problem. The amount of power and control Google has over the Internet is just scary. Google is the only company that manages to get me excited about cool stuff AND totally creep me out with each announcement they make.

    As for the need for code to remain closed to avoid abuse, I think open code would quickly evolve its own defense mechanisms against that sort of behavior, and decentralized search would make it harder for people to abuse ALL the search systems at once.

    At what point do we tell Google to let go of some of its power?

  39. Open Systems are not the same as Open Source, people.

    Read the article. It says Open Systems, Open Standards: use these always, without exception (and where there is no standard, create your own, and make it open for everyone else). Open Source: all developments are presumed to be open source-able until proven otherwise.

    This is the polar opposite of the way most large corporations guide their developers. In most large organisations, the standing orders are “whenever you develop something, always consider if it should be patented”.

  40. @Rob, read Google’s post again, they admit they’re a company and need to make money. Their point in the paragraph you quoted is valid, by keeping GMail/etc app closed they force competition to make something better from scratch, take Etherpad vs GDocs former Writely for example, if it was a stripped down version of the OSS GDocs they’d probably focus on features/something other than the base technology.

    Not all google acquisitions have a similar story/purpose, take reCaptcha for example, lol.

  41. The point isn’t whether Google is good or evil, or whether or not they should keep some of their stuff closed, which they do, and fight for a lot of stuff to be open, which they do.

    The point is, Google is trying to redefine “open” in a way that benefits Google, while ignoring the realities of “open” that Google doesn’t want open, because it interferes with how they make money as a corporation. That is the hypocrisy. If they were honest about open, they would say: “Open almost always wins. Except when it loses. Loses us money. Then closed always wins.”

  42. Ah, “The meaning of open”. I must admit, i’m a little confused by the term, as Google uses it.

    Do they mean “open” as in the Android Market, where users can not download the client or can not access it if their Android phone or device doesn’t come with it pre-installed – i assume that means the manufacturers paid Google for that right, or that Google means to get money from the future manufacturers?

    Or do they mean “open” as in sending C&D letters to developers that make their products better and more useful?

Comments are closed.